International Journal of Medical Engineering and Informatics (89 papers in press)
Hospitalization characteristics of Metabolic Syndrome patients
by Nimisha Patel, Riddhi Vyas, Shankar Srinivasan, Dinesh Mital
Abstract: Metabolic syndrome is a combination of disorders and in conjunction increases the risk of developing several chronic diseases. This study sought to determine the overall in hospitalization characteristics of metabolic syndrome and non-metabolic syndrome patients. This was a cross sectional study with descriptive analysis from Healthcare Cost and Utilization Project (HCUP) National Inpatient Sample (NIS) dataset from 2012 to 2014.Compared with Non-metabolic syndrome patients, metabolic syndrome patients length of stay was longer at mean 5.10 days versus mean 4.57 days for non-metabolic syndrome patients. Additionally, total in-hospital charges for metabolic syndrome patients was 30% higher than non-metabolic syndrome patients. Risk of developing metabolic syndrome in female was slightly elevated than in men. Having metabolic syndrome in white ethnic group was high and exhibited substantial differences among different ethnicity. Lower socioeconomic status patients were 37% more prevalent in having metabolic syndrome than the higher income patients.
Keywords: Metabolic syndrome; Non-metabolic syndrome; Hospital cost; Length of stay.
Diagnosing Angiographic Disease Status with the Aid of Deep Neural Network
by Jayakumari Damarla
Abstract: In this decade, one-third of all global deaths reason as cardiovascular diseases- a report from the World Health Organization (WHO). Early diagnosing conserves human lives from cardiovascular diseases, which is possible through computational techniques. This research intends to identify normal/abnormal conditions of heart diseases appropriately with the aid of the Artificial Intelligence (AI) technique. This research includes Deep Neural Network (DNN) to identify heart conditions adequately. It is evident from the investigation that DNN unveils 93.4% accuracy, which is proficient performance over other employed techniques. The performance of the research evaluates through nine-measures, where the DNN shows the superiority over contest techniques in all performance measures while predicting heart disease conditions.
Keywords: Artificial Intelligence (AI); Deep Neural Network (DNN); Angiographic Disease and Prediction/Classification/Diagnosing.
EFFICIENT TUMOUR DETECTION FROM BRAIN MR IMAGE WITH MORPHOLOGICAL PROCESSING AND CLASSIFICATION USING UNIFIED ALGORITHM
by G. Sethuram Rao, D. Vydeki
Abstract: Brain diseases caused due to malignant are the biggest concern among all the age groups. Studies show that almost 80% of death cases are reported due to presence of malignant tumour. Hence diagnosing brain tumour at an early stage would increase the survival rate. Magnetic resonance imaging (MRI) plays a major role in diagnosing tumours in human brain. However, it is considered to be a time consuming and tedious process which could lead to deviation in the opinion of radiologists. This has led to the development of computer-based automatic extraction of tumour cells from the images obtained by MRI. This paper proposes an efficient tumour detection mechanism from MR images using morphological processing and unified algorithm. A neural network that uses bounding boxes and associated class probabilities detects the
packets of tumour that exist in a full MR image. Simulated results of the proposed technique on the BRATS 2016 dataset show that a detection accuracy of 95.97% is achieved, while reducing the likelihood of false positives. This approach is compared with other detection methods such as DPM and R-CNN and the analysis proves that our method proposed outclasses the other detection methods.
Keywords: terms-magnetic resonance image; brain tumour; thresholding; histogram; segmentation; CLAHE; unified detection; malignant; benign.
Adaptive neuro-fuzzy-based attention deficit/hyperactivity disorder diagnostic system
by Anoop Kumar Singh, Deepti Kakkar, Tanu Wadhera, Rajneesh Rani
Abstract: The main purpose of this research paper is to develop a simple automated system for the accurate diagnosis of attention deficit/hyperactivity disorder (ADHD) using the adaptive neuro-fuzzy inference system (ANFIS). The designed diagnostic system has two stages-primary and secondary. In the primary stage, a hierarchical fuzzy-based short version of the gold diagnostic tool Connors scale has been implemented to evaluate the behavioural aspects in a fast and simple manner. The secondary stage targets the two main abilities of brain functionality-attention and perception. The determining traits were extracted from ERP components, especially the P300 wave, using peak amplitude and average latency rate. The proposed secondary diagnostic stage is based on Takagi-Sugeno fuzzy inference system and it integrates the features of both artificial neural network and fuzzy logic into a single framework. The system accuracy is 99.3% in classification, i.e., ADHD vs. Normal and 88.78% in severity level (normal/low, medium and high) of ADHD. Thus, the proposed model provides an adaptive and better alternative to ADHD diagnosis.
Keywords: attention deficit/hyperactivity disorder; ADHD; artificial neural network; fuzzy logic; backpropagation; ANFIS; neuro-fuzzy inference system; event-related potential; FIS; standalone fuzzy inference system.
MODELING AND ANALYSIS OF KNEE AND HIP JOINTS IN HUMAN BEING
by Bhaskar Kumar Madeti
Abstract: ABSTRACT: The present work aims at developing a representation of all the forces by first drawing the free body diagrams of the knee and hip joints. In order to do force analysis one needs to study knee and hip anatomy. With the aid of MRI Scan data, the moments of the forces are computed so as to solve the equilibrium equations. A 3-dimensional (3-D) finite element analysis is generated to represent the real world situation as closely as possible. The accuracy is improved using image processing commercial software on Magnetic Resonance Imaging (MRI) and Computed Tomography (CT) scans. The analysis was conducted for body weights of 600 N, 1000 N and 1500 N for each of the possible postures during various activities. One important recommendation that can be made from present work is that in order to find proper replacement of human knee and hip joint, one needs to collect CT scan slices and then create 3D models performing F.E analysis by selecting the persons weight. In other words the selected implant must be customized for the patients weight, rather than making the choice by rule of thumb as in common practice in hospitals today.
Keywords: Keywords: CT scan; MRI Scan; 3D model; FE analysis; Knee; Hip.
STUDY OF DEMOGRAPHIC AND RISK FACTORS ASSOCIATED WITH LIVER DISEASES
by Disha Sheth (Kothari), Riddhi Vyas, Shankar Srinivasan
Abstract: ABSTRACT:Liver diseases can be diagnosed by interpreting enzyme abnormality pattern and patient characteristics. Despite growing evidence suggesting different causes, there is a need to explore the risk factors which lead to liver diseases in different population. NHANES Data (2015-2016) was investigated for gender, age, race and country of birth in patients with high liver enzymes values. Different elements were also studied for significant liver disease risk based on odds ratio, 95% confidence interval and relative risk using Fisher Exact Test. Results showed US-born, Non-Hispanic white young males had high values for liver enzymes, demonstrating greater risk in such population for liver diseases. Odds ratio (<1.0); P-value of significance (<0.0001) indicated negligible risk associated with all elements - iron overload, diabetes, smoking, alcohol, blood pressure, total cholesterol, obesity, total protein, albumin and total calcium. Validation studies were also performed using NHANES data (2013-2014), authenticating the results obtained.
Keywords: Liver Diseases; Demographic; Odds Ratio; Relative Risk; Fisher Exact Test.
A novel UWB compact elliptical-patch antenna for
early detection of breast cancer in women with high mammographic density
by Amber Khan, Mainuddin , Moin Uddin, Parikshit Vasisht
Abstract: Microwave imaging is one of the emerging technologies for early detection of breast cancer among women having dense mammographic densities. One of the critical and valuable components of an accurate, effective and compact, involving minimum risk microwave imaging system for early breast cancer detection is an ultra wideband (UWB) antenna. A novel, compact elliptical UWB microwave antenna is presented in this research article that might be suitable for early breast cancer detection. The simulation of antenna structure is carried out using HFSS13 FEM-based EM software. The simulation results yield better UWB response. The antenna structure provides a wide practical fractional bandwidth of more than 156%. A significant performance
factor of the proposed antenna is its ability to provide sufficient gain level for
short distance communication. Thus, the proposed antenna is a strong candidate for design and development of microwave imaging system for early detection of breast cancer among women with dense mammographic densities.
Keywords: ultra wideband; UWB; wireless body area network; WBAN; elliptical-patch; quality of service; QoS; UWB sensors; breast cancer.
Diagnosing Lung Cancer with the aid of BPN in associate with AFSO-EA
by Rahul Shreyas, Gopika Kumari
Abstract: This work aims at identifying lung cancer into various classes of carcinomas or as a normal-lung, with the aid of an artificial neural network classifier. One thousand input attributes obtained from multiple modality images of the lung, and four output classes are defined. Existing works in the area choose to maximize the accuracy of classification as the primary goal of their researches. However, the reduction of the process complexity has been a relatively untouched area. Few of the available works have tried out the possibilities of reducing the number of input features. This work aims at modeling an optimum Back Propagation Network (BPN) model, by reducing the input feature count and by optimizing the number of neurons in each layer of the BPN classifier without compromising the accuracy. This work incorporates Artificial Fish Swarm Optimization (AFSO) and Evolutionary Algorithm (EA) and proposes a hybrid AFSO-EA for reducing the input feature set. This work also configures a BPN model, where the number of neurons in each hidden layer is optimized using the same hybrid method. The investigation results reveal that the proposed hybrid AFSO-EA technique generates a BPN model, which can achieve 97.5% classification accuracy, with much less computational overhead, than the existing methods.
Keywords: Lung cancer; Backpropagation network (BPN); Levenberg-Marquardt (LM); Artificial Fish Swarm Optimization (AFSO); Evolutionary Algorithm (EA) and hybrid Artificial Fish Swarm Optimization - Evolutionary Algorithm (AFSO-EA).
Operating room scheduling 2019 survey
by Maha TOUB, Omar SOUISSI, Said ACHCHAB
Abstract: Numerous optimization problems in Healthcare have been approachedrnby researchers over the last three to four decades. Hospital logistics - organized and structured to secure patient satisfaction in terms of quality, quantity, time, security and least cost - forms part of the quest for global performance. We provide herein a review of recent study and applications of Operations Research in Healthcare. In particular, we survey work on optimization problems, focusing on the planning and scheduling of operating rooms. The latter is a highly strategic place within the hospital as it requires key medical competence and according to related works surgical sector expenditure represents nearly a third of a hospitalsrnbudget. We analyze recent research on operating room planning and schedulingrnfrom 2008 to 2019; our evaluation is based on patient characteristics, performance measurement, the solution techniques used in the research and the applicability of the research to real life cases. The searches were based on Pubmed, Web of science, sciencedirect and google scholar databases.
Keywords: Operation Research; Healthcare; Operating room; Scheduling;rnPlanning; Optimization; Surgery scheduling; Literature review.
Detection of Abnormality in Breast Thermograms using Canny edge detection algorithm for Thermography Images
by Kumod Gupta, Ritu Vijay, Pallavi Pahadiya
Abstract: Currently research towards cancer is gaining fast attention, as methods to cure
cancer are a holy grail. Among many potential techniques, breast cancer thermography
techniques may come up in saving many lives in the future. The purpose of this paper is
to diagnose breast cancer at preliminary stage using infrared breast thermography. In the
first approach, the thermography image is acquired and conclusions are drawn on the
basis of their symmetry using the histogram, is not appropriate to take decision for
practitioner. In the second approach image is processed and apply algorithms to get good
result. Further, it also helps us to explore those statistical features that effectively
distinguish healthy breast thermograms from that of the thermograms caused by a disease.
Finally, graphical representation of the data corresponding to statistical features for both
the left and right breast of the Healthy and sick Patients breast thermogram has been
made in this paper. The mammography report is carefully examined and compared to
signify any abnormality. The values obtained from asymmetric analysis based on the
abnormality detection system are 94.44% of Sensitivity, 83.33% of Specificity and
88.88% accuracy. This presented work is fruitful for the medical practitioner in early
detect breast cancer.
Keywords: Infrared Radiation thermograms (IRT); Mammography Images; Feature Extraction; Malignant; Benign; Region of interest (ROI);.
An automatic ECG arrhythmia diagnosis system
using support vector machines optimised with GOA
and entropy-based feature selection procedure
by Abdullah Jafari Chashmi, Mehdi Chehel Amirani
Abstract: Primary recognition of heart diseases by exploiting computer-aided diagnosis (CAD) machines, decreases the vast rate of fatality among cardiac patients. Recognition of heart abnormalities is a staggering task because the low changes in ECG signals may not be exactly specified with eyesight. In this paper, an efficient combination classification model using grasshopper optimisation algorithm (GOA) and support vector machines (SVMs) called GOA-SVM for ECG arrhythmia diagnosis is proposed. In this approach, the combination of discrete wavelet transform and higher-order statistics is used to feature extraction and the entropy-based feature selection method. The proposed method has been compared with PSO-SVMs and SVM-RBF kernel function for classifying the five classes of heartbeat categories. Our proposed system is able to classify the arrhythmia classes with high accuracy (99.66%). The simulation results show that classification accuracy in SVM-GOA method is better than SVM-RBF and neural network classifier.
Keywords: ECG; classification; entropy; grasshopper optimisation algorithm; GOA; higher order statistics.
Protection of Encrypted Medical Image using Consent based Access Control
by Mancy Lovidhas, Maria Celestin Vigila S
Abstract: An outline which defends tolerant details during facts transfer be necessary for medical management systems. On the way to attain safety and confidentiality for facts transfer, a consent based access control system was proposed. It grants the agreement by distributing token to the data client, where the permission can only be created by official client. Thus the information stored inside the data centre can be accessed only when the data requester has the token, which is similar to the token already present inside the data centre. If the confirmation of data centre is valid, the data requester can access the original information of the user. Eventually, the user will be notified by the data centre to deserve that there is no misuse outside consent. The anticipated consent based access control method is compared with existing methods to achieve less time utilization and low computational overhead.
Keywords: Consent ; Authorization ; Data Requester ; Data Center ; Data Provider.
An Innovative Hearing-Impaired Assistant with Sound Localization and Speech to Text Application
by Balaganesh Baskar, B.V. Damodar, R. Dharmesh, K.R. Tharunkarthik, K.V. Shriram
Abstract: According to the statistics of World Health Organization (WHO), there are about 466 Million people (About 5% of the total population worldwide) are hearing impaired and 34 million of them are children. It is estimated that by the 2050, there will be almost 900 Million people suffering from hearing disability. In India, there are 63 Million people with hearing impairment. Hearing aid prices range from ?20,000 for a basic device to ?2,50,000 for a premium hearing aid. People with hearing disabilities should not have to spend so much money to enable a sense that normal people take for granted. One of the main problems faced by a deaf person is that they find it difficult to have casual conversations because it is hard for them to follow what others are speaking. This can be addressed simply using a mobile application. We present a frugal and affordable system that could show the direction of the speaker along with the speech in text format in real-time. This can be achieved by Sound-Localization and Speech-to-Text conversion. Sound-Localization is a technique used to identify the direction of the source of the sound. There are Speech-to-Text tools that can generate text from a speech in real-time.
Keywords: Sound Localization; Android Application; Simple Conversation Application; Speech to Text.
An efficient AR modeling based Electrocardiogram signal analysis for Health Informatics
by VARUN GUPTA
Abstract: Today health informatics not only require correct, but also timely diagnosis much before the occurrence of critical stage of the underlying disease. Electrocardiogram (ECG) is one such non-invasive diagnostic tool to establish an efficient computer-aided diagnosis (CAD) system. In this paper, autoregressive (AR) modeling is proposed that is an efficient technique to process ECG signals by estimating its coefficients. In this paper, two parameters viz. Atrial Tachycardia (AT), and Premature Atrial Contractions (PAC) are considered for evaluating the performance of the proposed methodology for a total of 17 recordings (6 real time and 11 from MIT-BIH arrhythmia database). As compared to K-Nearest Neighbor (KNN) and Principal Component Analysis (PCA) with AR modeling [ also known as Yule-Walker (YW) and Burg method], KNN classifier coupled with Burg method (i.e. Burg+KNN) yielded good results at model order 9. A sensitivity (S_e) of 99.95%, specificity(Sp or PPV) of 99.97%, detection error rate (DER) of 0.071%, accuracy(Acc) of 99.93% and mean time discrepancy (MTD) of 0.557 msec are obtained. Consistent higher values of all the performance parameters can lead to the development of an autonomous CAD tool for timely detection of heart diseases as required in health informatics.
Keywords: ECG; AR coefficients; Atrial Tachycardia; Premature Atrial Contractions; KNN classifier; PCA classifier; Yule-Walker; Burg method.
A supervised learning model for medical appointments no-show management
by Inês Ferreira, André Vasconcelos
Abstract: A no-show is a phenomenon that leads to an efficiency decrease in various sectors, including in the healthcare sector. This research proposes the usage of supervised learning techniques to predict medical appointments no-shows occurrence and to find patient replacements to fulfil last-minute vacancy slots. The prediction is performed using a classification algorithm that computes the probability of no-show for each patient based on features that have shown to influence his or her decision, such as the waiting time, the day of the appointment and the number of previous no-shows, among others. The features are extracted from two distinct healthcare datasets. In order to select the most suitable classification algorithm, a ten-fold cross-validation is used to perform a comparative analysis among the most used algorithms applicable to this type of classification problems. The gradient boosting algorithm proved to have the best performance in estimating no-shows.
Keywords: no-show; healthcare; supervised learning; classification algorithms; cross-validation.
Sentiment Analysis of an Epidemic: A case of Nipah Virus in India
by Jayan V, Sreejith Alathur, Rajesh R. Pai
Abstract: Data in social media and other news media can have an impact on the decision-making process of the Government and the citizen if properly examined. The mode and pace of dissemination in both media leads to an increase in the delivery of misinformation. This affects the economy of the country and people's mental health. The government must formulate the required measures to counter the proliferation of fake messages and disinformation in the media, which would otherwise lead to an unnecessary burden. Regulation of health communication during the period of epidemic is important, as it has an effect on the mental health of users of the media. The study assesses the emotions of health communication in social media and online news media in the context of the Nipah Epidemic in India during 2018.
Keywords: Fake News; Nipah; Sentiment Analysis; Social media; psycho-linguistics; neuro-linguistics; misinformation; depression; anxiety.
Hierarchical cluster analysis of the morbidity and mortality of COVID-19 across 206 countries, territories and areas
by Donald Douglas Atsa'am, Ruth Wario
Abstract: This research deployed the agglomerative hierarchical clustering to extract clusters from the coronavirus disease 2019 (COVID-19) data based on the morbidity and mortality of the novel virus across 206 countries, territories and areas. As of 2 April, 2020, a total of 896,475 confirmed cases were reported across the world. Three clusters were extracted from the data on the bases of morbidity and mortality of COVID-19. These include: low-confirmed-cases, low-new-cases, low-deaths and low-new-deaths countries [cluster 1]; medium-confirmed-cases, low-new-cases, medium-deaths, and medium-new-deaths countries [cluster 2]; high-confirmed-cases, high-newcases, high-deaths, and high-new-deaths countries [cluster 3]. It is recommended that, to contain the pandemic, countries within a cluster should cooperate, share information and learn from mistakes or strategies (as the case may be) of the countries in other clusters. Among other benefits, this can prevent countries within the low-confirmed-cases cluster from progressing to the high-confirmed-cases cluste
Keywords: COVID-19; morbidity; mortality; hierarchical clustering; data mining.
A new approach based on for controlling the joint movement of drop foot patients
by Mina Lagzian, S. Ehsan Razavi, Hamid Reza Kobravi
Abstract: Stepping is one of significant functions, which needs an appropriate coordination between various joints to be accomplished properly. Drop foot is a gait abnormality that the harmony between joints is disturbed. In this paper, proposed a new fuzzy control model for controlling joint movement of drop foot patients. This method has two advantages over pervious ones. The first one is based on identify kinematic pattern not just statistical works. The second is independent of any mathematical models and formulas. The controlling method is based on identification of both stable and unstable manifolds of basin attraction of a healthy person in order to, how to properly move his/her defective leg. The results indicate that using the proposed fuzzy controlling approach, has lower computations and good convergence.
Keywords: stepping procedure; drop foot; gate analysis; stable and unstable manifolds; absorption platform; saddle points; fuzzy control.
A secure and intelligent real-time health monitoring system for remote cardiac patients
by Maroua Ahmid, Okba Kazar, Laid Kahloul
Abstract: In this paper, we propose an intelligent and secure internet of things approach for the healthcare system that monitors the patient heart rate in real-time and from any place. Thanks to the agent, the proposed system can predict the critical condition before it even happens and takes fast and apt decisions in an emergency case. Based on the experimentation, the proposed system is convenient, reliable, and ensures data security at a low cost. The proposed algorithm outperforms other algorithms regarding the systems operational efficiency. It is more suitable for devices with power, storage, and processing limitations, such as in IoT devices. Also, agents are the better current technologies for heterogeneous and distributed systems, such as the internet of things. Moreover, this approach scalability makes it suitable for a
broad range of IoT environments, including smart homes, smart cities, dynamic and large-area networks, etc.
Keywords: healthcare; agent; ECC EIGamal; remote monitoring; cloud; internet of things; IoT.
Virtual Reality-Based Real-Time Solution for Children with Learning Disabilities and Slow Learners An Innovative Attempt.
by K.V. Shriram, Pranav B, Saravanan G, Merin K. John, Athira Sasidharan
Abstract: Autism Spectrum Disorder (ASD) is a developmental disorder which can be characterized by social and communication impairments, slow learning, combined with limited interests and repetitive behaviors. It affects as many as one in 59 children and is more prevalent in boys with one in 38 diagnosed with Autism Spectrum Disorder. But due to the social stigma associated with mental health and psychological issues, especially in countries like India, most cases go unreported or symptoms ignored. This project is an attempt to help address this issue by providing a means to assist in diagnosing ASD using telemedicine and also to provide an interactive and effective means of learning for children diagnosed with ASD or children with slight learning disabilities. The system features games that have been proven to be effective with children diagnosed with ASD. The virtual reality which is being speculated to be a powerful tool in helping children with learning difficulties has been used to enhance the effectiveness of these games. The concept of dynamic difficulty is also integrated into the game in order to increase or decrease the challenge of each level depending on the performance of the child which further increases the effectiveness of the game.
Keywords: Autism; Slow Learners; Learning Difficulties; 3D; game; Virtual Reality; Hand Tracking; Interactive; Dynamic difficulty; Adaptive levels; Gesture detection; Leap Motion sensor;.
An Innovative Deep Learning Approach for COVID 19 Detection with X-Ray Images and Infected User tracking through Blockchain
by Vimal Kumar, Shriram K Vasudevan, Nitin Dantu
Abstract: The COVID-19 pandemic has shocked the globe with an enormous number of people infected and a large death toll across several nations. Many people lost their loved ones and 350,000 death toll passed globally. By this time more than five million people have been affected. A deadly virus has many victims but no country could stand out when it comes to producing a vaccine. The virus is so dangerous that it spreads rapidly through human contact and a person who is infected will infect around 600 people a month. It is so fast that more than 50,000 people are affected in one day in some countries and more than 1,000 people die in one day. The present situation is so bleak, and if not contained by social distance, it can get even worse. There are many patients but not enough doctors and hospitals to treat them as the infection grows exponentially. No doctor can examine Chest X-ray in thousands and have fast turnaround. We want to create a solution to reduce the workload on doctors, to easily determine whether a Chest X-ray pneumonia is due to coronavirus or not, so that the rapid spread can be controlled and proper cure could be given to patients. Here we also add the distributed ledger technology called blockchain, which helps in monitoring the patient health data and thus it helps in having the complete history of the patient.
Keywords: Covid 19; Covid 19 with Deep Learning; Deep Learning; Blockchain for Covid 19; Covid 19 with XRAY;.
Analysis of Dermal Activity and Skin Images for Diabetic Kidney Disease
by Valli MN, Sudha Singaram, Kalpana Ramakrishnan, Soundararajan Periasamy
Abstract: Any electrical input to skin, changes the ion concentration in sweat, leading to variations in electro dermal activity (EDA) and hence in skin conductivity. Structurally, the pores and connecting tissues contribute to skin texture. Diabetes leads to micro-vascular complications thus affecting the innervations of C-nerve fibers, thereby skin conductivity and micro texture also changes. Diabetic kidney disease (DKD) is another condition under which hydration level and urea in serum and sweat varies leading to dermal changes. Therefore EDA and microscopic-images are acquired from volunteers catering to normal, diabetic and DKD. Features are extracted after convolving EDA signals with Morlet-wavelet and pre-processing the micro texture image for hair removal and enhancement. An expert system is designed to take these features as input and for broad classification. Result of this study demonstrates the influence of serum urea on skin conductivity and texture, thereby enabling skin based method of diagnosing diabetics and DKD.
Keywords: Electro dermal activity; kidney; feature extraction; artificial neural network.
APPROACHES AND CHALLENGES TO SECURE HEALTH DATA
by Patricia Whitley, Hossain Shahriar, Sweta Sneha
Abstract: As the volume of health data being generated and stored massively, the number of data breaches are also increasing causing concerns among patients and healthcare providers on how to protect data better. This article explores blockchain, machine learning and artificial intelligence as possible technologies to secure healthcare data and some challenges when incorporating them to mitigate against data breaches. The paper also discusses a discussion of the issues surrounding the security of health data and improvement.
Keywords: EHR; Data Security; Blockchain; Machine Learning; Artificial Intelligence.
Impact of wireless technologies on public health: a literature review
by Antonio Conduce, Daniela Di Sciacca, Sergio Sbrenni
Abstract: Introduction: The aim of this review is to evaluate the impact wireless technologies have on the public health in terms of patient safety, quality of care and cost savings. Methods: A systematic review was performed. We ended up analysing 76 papers and found the main applications of wireless technologies on public health in the literature. Results were organized in four different categories, one being a subsequent refinement of the previous one. Results: This study identify and analyses the risks and benefits on public health, highlighting strengths and opportunities, especially for patients in prehospital stage. The most relevant benefits identified are: improving outcomes in time-dependent pathologies and reducing management costs. Conclusion: The adoption of wireless technologies in healthcare is still in a trial stage. A careful evaluation of their impact on the quality and sustainability of health services has to be performed in order to obtain the final approval.
Keywords: Emergency Medicine; Ambulances; Quality of Care; Wireless technology; Equipment and Supplies; Telemedicine; Technology Assessment; Biomedical; Electronic Health Records; Review.
THE EFFECT OF SOCIOECONOMIC FACTORS ON HEALTH-RELATED QUALITY OF LIFE AMONG ADULTS WITH DEPRESSIVE DISORDER IN THE UNITED STATES
by Nesren Farhah, Shankar Srinivasan, Dinesh Mital, Frederick Coffman
Abstract: Using data from the Behavioral Risk Factor Surveillance System (BRFSS) a study was conducted to determine the effect of socioeconomic factors of education level, marital status, employment status, and income level on the HRQOL outcomes of activity limitation, physical health, and mental health among adults with depressive disorder in the United States. A greater number of adults with high income level, high education level and married were depression free compared to those with low incomes (39.17% vs 6.49%), low education level (30.46% vs 5.8%), and being single (45.35% vs 8.35%). Also, those with depressive disorder suffered greater physical health problems (11.02% vs 7.93%) and mental health problems (12.58% vs 6.26%).
Keywords: Depressive disorder; socioeconomic factors; Health-related quality of life; mental health; physical health; activity limitation.
MSCs-released TGF?1generate CD4+CD25+Foxp3+ expression in T-reg cells of Human SLE PBMC
by Dewi Masyithah Darlan, Delfitri Munir, Agung Putra, Nelva Karmila Jusuf
Abstract: Regulatory T-cell (Treg) defects may cause autoreactivity of both T and B cells leading to autoimmune disease, including in Systemic lupus erythematosus (SLE) disease. Those defects were characterized by decreased expression of CD4, CD25, and FoxP3, thus restoring the Treg expression can reverse autoimmunity into immune tolerance into a normal immune response. Mesenchymal stem cells (MSCs) have immunomodulatory properties to control inflammation milieu, including in SLE inflammation by releasing TGF?1, IL-10, and PGE2, thus MSCs can generate Treg cells. However, the regulation of Treg by MSCs-released TGF?1 in human SLE remains unclear. This study aims to analyze the role of MSCs-released TGF?1 in generating CD4+, CD25+, Foxp3+expression in T-reg cells of human SLE PMBCs. This study used a post-test control group design using the co-culture of PBMCs from SLE patient and human umbilical cord MSCs (hUC-MSC) as the subject. This study was divided into 5 groups; sham, control, and treatment group treated by co-cultured hUC-MSC to PBMCs with ratio 1:1 (T1), 1:25 (T2), and 1:50 (T3) for 72 hours incubation, respectively. The expression of T-reg was assessed by flow cytometry assay, whereas the TGF?1 using Cytometric Bead Array (CBA).This study showed a significant increase in Treg cell expression (P
Keywords: MSCs; TGF?; CD4+CD25+Foxp3+; T-reg; SLE disease.
An ensemble framework-stacking and feature selection technique for detection of breast cancer
by Vikas Chaurasia, Saurabh Pal
Abstract: Breast cancer is the second most common cancer in women worldwide. The machine learning (ML) method is a modern and accurate technique that researchers have recently applied to predict and diagnose breast cancer. In this research article, we developed stack-based ensemble techniques and feature selection methods for the comprehensive performance of the algorithm and comparative analysis of breast cancer datasets with reduced attributes and all attributes. This article uses five-feature selection technique because it affects the overall performance of the model. After applying feature selection method, now we have dataset with reduced features as well as all features. We implemented logistic regression on a dataset with all features and a dataset with reduced features. Finally, we see that the dataset with reduced
features have got improved accuracy.
Keywords: breast cancer; k-nearest neighbour; KNN; perceptron; stacking; machine learning; feature selection; algorithm; ensemble techniques; logistic regression; sub-models.
DEPRESSION CLASSIFICATION AND RECOGNITION BY GRAPH-BASED FEATURES OF EEG SIGNALS
by Ahad Mokhtarpour, Faezeh Bashiri
Abstract: Major depressive disorder(MDD) is one of the main subjects in world health so its diagnosis is important for researchers. Electroencephalography(EEG) is one of effective tools in brain psychological disorders diagnosis which any change in brain function is reflected in signals. By EEG signal analyzing, some disorders like MDD can be recognized. In this paper EEG signals are firstly mapped to four different visibility graphs and several features are extracted from each graphs. Then feature numbers are reduced by principal component analysis(PCA) and depressive and normal classification is done by support vector machine(SVM). In this paper, classification results by combining all four graph features are compared with each graph features individually and the results show that by combining features lower classification error and better accuracy is achieved. The classification accuracy of depression classification by mixed features is 100 percent which means the proposed method can classify all of them correctly.
Keywords: Electroencephalography; major depression disorder; visibility graphs; support vector machine.
Effective Utilization of Multi Median Variance-Independent Component Analysis on Medical Image Denoising
by Arathi Thiruvoth, Rahul Chingamtotatil
Abstract: Image denoising is a significant pre-processing technique that plays a vital role in medical image processing. Image denoising is the process of removing noise from an image and is a trade-off between noise removal and preservation of significant image details. This paper encloses a sparse representation based denoising technique called Multi Median Variance-Independent Component Analysis (MMV-ICA). Investigation evident, the incorporation of MMV ICA reveals superior denoising results over contest techniques under various noise attacks and noise level conditions. The proposed denoising algorithm based on sparse and redundant representations over learned dictionaries. The dictionary is trained using the corrupted image, and after that, the dictionary is adapted to achieve sparse signal representations. MMV-ICA algorithm presented in this paper makes use of a patch-based dictionary creation method. The paper presents the results of the MMV-ICA denoising technique, which are found to be in par with the existing sparse based denoising methods.
Keywords: Image Denoising; Sparse Representation and Multi Median Variance-Independent Component Analysis.
Clinical decision support system for early diagnosis and intervention in multiple sclerosis
by Jojy Cheriyan, Dinesh P. Mital, Shankar Srinivasan, Riddhi Vyas
Abstract: Multiple sclerosis is one of the most common neurological disorders causing disability among young adults in North America and Europe. It is an incurable and non-communicable disease, debilitated by physical and mental impairments. Recently the incidence of MS reported in the USA doubled, compared to the past estimates. The average period to diagnose MS still ranges from six months to three years despite all advances in diagnostics and therapeutics. Studies suggest that early diagnosis and intervention can delay the progression of the disease and improve the quality of life. We have done a review of literature followed by a quasi-experimental approach to collect quantifiable predictors and analyse the trending incidence of MS in the USA. We conclude that no CDSS exists to date to diagnose MS early. Predictors are
available to design a clinical decision support tool for multiple sclerosis that
can help clinicians in the early diagnosis and intervention.
Keywords: multiple sclerosis; MS; clinical decision support system; CDSS; national inpatient sample; NIS; healthcare cost and utilisation project; HCUP.
Classification and signal processing analysis
Of The pathological electromyogram signal (EMG)
by Mokdad Aicha, Debbal Sidi Mohammed El Amine, Meziani Fadia
Abstract: The objective of this ongoing study is to introduce electromyography signal (EMG) in time-frequency representation (TFR) applying spectrogram with optimized window size where four features were extracted. In order to qualify or not the capability of spectrogram features in separating healthy and amyotrophic lateral sclerosis (ALS) pathology, three useful classifiers namely support vector machine (SVM), linear discriminate analysis (LDA), K-Nearest Neighbor (KNN) are implemented to classify EMG signals.AS result, spectrogram with optimized window size (512 ms) and SVM based on Radial basis function (RBF-SVM) presents the highest classification accuracy of 92.3% Followed by LDA and KNN with classification accuracy of 90.86% and 83.3% respectively, where the optimized window size of 256 ms is more appropriate. Also, the proposed TFR is able to show the nonstationary variations of sEMG signals. the features exhibit statistically significant difference in the muscle healthy and neuropathic conditions. The combination of RBF based SVM is found to be most accurate (92.3% accuracy) in classifying the conditions with the extracted features based on spectrogram.
Keywords: Amyotrophic lateral sclerosis; spectrogram; classify; support vector machine; linear discriminate
Study of Novel COVID-19 Data using Graph Energy Centrality: A Soft Computing Approach
by Mahadevi S., Shyam S. Kamath, Pushparaj Shetty D.
Abstract: The propagation of the new pandemic COVID-19 is more likely linked to human social relations and activities. A Social Network can be used to describe these human relationships and activities. Understanding the dynamic properties of disease dissemination through diverse Social Networks is critical for effective and efficient infection prevention and control. With the frequent emergence and spread of infectious diseases and their impact on large areas of the population, there is growing interest in modelling these complex epidemic behavior. Such an approach could provide a stronger decision-making method to tackle and control disease. In this paper, a transmission network is developed using the South Korean data, and the study of the network is carried out using Graph Energy Centrality. This measure of centrality allows us to recognize the primary cause of the spread of the virus within the established network by ranking the nodes of the network based on graph energy. The identified primary cause can then be isolated, which can prevent further spread of infection. We have also considered the Novel_Corona_Virus_2019_Dataset from Johns Hopkins University to analyse epidemiological data around the world.
Keywords: Coronavirus; SARS-CoV-2; Centrality Measures; Graph Energy; Data Analysis; Visualization; Social Network Analysis.
Early diagnosis of coronary artery disease by SVM, decision tree algorithms and ensemble methods
by Marziye Narangifard, Hooman Tahayori, Hamid Reza Ghaedsharaf, Mehrdad Tirandazian
Abstract: Heart diseases are one of the main causes of death around the world. The most reliable method for heart disease diagnosis is angiography, which is costly, invasive and has the risk of death. This study applies variations of decision tree (DT), support vector machine (SVM) and voting algorithms to construct a heart disease diagnosis predictive model. We show that integrating medical knowledge and statistical knowledge as well as fine tuning the parameters of the used models lead to more effective heart disease diagnosis models. We use two methods for implementing the proposed model. The
obtained results in both methods show that voting algorithm and Random Forest outperform other methods. Moreover, the achieved accuracies show improvements over other existing methods.
Keywords: data mining; machine learning; decision tree; support vector machine; SVM; voting; random forest; forest PA; heart disease; UCI dataset.
Design of protective vessel and irrigation system for an organ-on-chip device
by Esmeralda Zuñiga-Aguilar, O. Ramírez-Fernández, Adeodato-Israel Botello-Arredondo
Abstract: New devices have been in development in the biomedical engineering field which allow to mimic several physiological processes at once or individually. The present work introduces a design and computational simulation of the nutrient irrigation system, as well as the rapid prototyping of the protective vessel of an organ-on-chip (OOC) device as a way to manipulate and transport the system easily as a whole while maintaining the proper irrigation conditions in the media. The device was generated with the computer-aided design (CAD) software, SolidWorks and the irrigation of the
system was performed with the aid of SolidWorks Flow Simulation module. The components of the presented OOC system were manufactured by 3D printing and by using the stereolithography technique. The results showed the flow velocity fields with values in the rage of 0.1830 m/s in the zone were the OOC is located, which indicates would allow a proper irrigation of nutrients to the cells in the chip. The proposed design of the OOC device as a whole, demonstrated to be an adequate storage and handling system for the OOC, in addition of providing a continuous irrigation of the medium.
Keywords: organ-on-chip; OOC; protective vessel; computer-aided design; CAD; irrigation system; 3D printing; flow simulation; biomedical device.
A neural network model for preeclampsia prediction based on risk factors
by Masoumeh Mirzamoradi, Atefeh Ebrahimi, Ali Ameri, Masoumeh Abaspour, Hamid Mokhtari Torshizi
Abstract: This study proposes a risk factor-based neural network model for preeclampsia prediction during the second trimester of pregnancy. A total of 320 women giving birth (160 normal delivery, 160 with preeclampsia) at Mahdieh Gynecology Hospital during 2018-2019, were inquired for 13 risk factors. Data from 85% of the subjects (selected randomly) were employed to train the network, and data from the remaining subjects were used to test the performance of the model. This process was repeated 100 times and the average results were determined. The proposed model achieved an accuracy of 83% in classifying the subjects into normal and preeclampsia classes, based on the risk factors input data, with a sensitivity of 83% and a specificity of 82%.
Keywords: Artificial Neural Network; Prediction; Preeclampsia.
Developing hybrid fuzzy model for predicting severity of end organ damage of the anatomical zones of the lower extremities
by Nikolay Aleexevich Korenevskiy, Alexander Vladimirovich Bykov, Riad Taha Al-Kasasbeh, Altyn Amanzholovna Aikeyeva, Sofya Nikolaevna Rodionova, Ilyash Maksim, Ashraf Adel Shaqadan
Abstract: Treatment of ischemic disease of the lower extremities is challenged by uncertainty and complex correlations among factors. Also, early detection and treatment have valuable effect on saving patients extremities from amputations caused by chronic obliterating diseases of the arteries of the lower extremities. We develop a hybrid fuzzy logic rules model to classify severity of ischemic lesions of the lower extremities so physicians can choose the prevention and treatment course. We use health indicators including rheological indicators, volumetric blood flow velocity, along anatomical zones of the lower extremities and regional systolic blood pressure. The model gives following classes: mild, moderate, severe and critical severity of ischemic damage. Medical experts judgement is integrated used in developing fuzzy rules. The proposed decision rule classification model exceeds 0.95. This model shows increasing necrobiotic changes in the ischemic extremities which helps preserves organ from amputation.
Keywords: critical ischemia; lower extremities; vascular blood filling; hemostasis; ischemic injury; fuzzy model.
COVID-19 detection through convolutional neural networks and chest X-ray images
by K. Venkata Subbareddy, L. Nirmala Devi
Abstract: To break the chain of COVID-19, a powerful and fast screening system is required which identifies the COVID-19 affected cases quickly such that the appropriate measures like Quarantine or treatment can be taken. The traditional Genetics assisted chain reaction test is found to have significant misclassification rate followed by more time consuming. To solve this problem, in this paper we have introduced a new model for COVID-19 detection based on Chest X-Ray (CXR) Images and Convolutional Neural Networks (CNNs). The proposed model is an automatic detection model which considers the CXR image as input and performs an in-depth analysis to discover the COVID-19. The proposed CNN model is a very simple and effective which is composed of five convolutional layers and three pooling layers. Every convolutional layer has different sized filters and different number of filters, which extracts all the possible features from CXR image. Simulation experiments are conducted over a newly constructed dataset based on the publicly available CXR (both COVID-19 and Non-CVOID-19) images. Simulation is done under two phases; 3-class and 2-class and obtained an average accuracy of 92.22% and 94.44% respectively. Thus the average accuracy is measured as 93.33%
Keywords: COVID-19; deep learning; convolutional neural network; CNN; CXR images; accuracy.
Melanoma classification by 3D colour-texture feature and neural network with improved computational complexity using PCA
by Mohd Firoz Warsi, Ruqaiya Khanam, Usha Chauhan, Suraj Kamya
Abstract: The most severe kind of skin cancer is malignant melanoma. It can grow anywhere on the body. Its exact cause is still unclear but typically it is caused by ultraviolet exposure from sun or tanning beds. Its detection plays a very significant role because if detected early then it is curable, before the spread has begun. In this paper a computationally improved [using principal component analysis (PCA)] feature extraction method named 3D colour texture feature (CTF) is represented which is well discriminative. For classification of melanoma from dermoscopic images, a comparison of different types of machine-learning classification algorithms is evaluated, out of which back propagation neural network (NN) classifier outperforms all other and produce best results, i.e., accuracy = 98.5%, sensitivity = 99.4%, specificity = 95.0%. Obtained results are even better than benchmarking results of PH2 dataset. Comparisons of results with other similar novel works are also discussed.
Keywords: melanoma; colour texture feature; CTF; dermoscopic image; neural network classifier; PCA; PH2; skin cancer.
Supervised classification approach for cervical cancer detection using Pap smear images
by Pallavi V. Mulmule, Rajendra D. Kanphade
Abstract: Cervical cancer is found in women and is the global life threatening problem. Papanicolaou test is the well-known technique used for diagnosing the cancer at the early stage. However, the pathological screening is manual, tedious and time consuming process. Therefore, the proposed method employs adaptive fuzzy k means clustering to segment the cell containing nucleus and cytoplasm from the unwanted background from the pathological Pap smear image. Thereafter, the 40 features are extracted from the segmented images based on the shape, size, intensity, orientation, colour, energy and entropy of nucleus and cytoplasm individually. Finally, supervised classification approach utilising multilayer perceptron with three kernels and support vector machine with five different kernels as the classifiers to predict the cancerous cells. The classifier is trained and tested on benchmark database with 280 Pap smear images. The performance of these two classifiers are evaluated and found that the MLP classifier with hyperbolic tangent activation function outperforms in all the performance criterias as compared to SVM classifier with classification accuracy of 97.14%, sensitivity of 98%, Specificity of 95% and positive predictive value of 98%.
Keywords: cervical cancer; Pap smear stain; pathological images; radial basis function; RBF; multi-layer perceptron; neural network.
MRI DENOISING: A SPARSE ICA BASED DICTIONARY LEARNING APPROACH
by Arathi Thiruvoth, Rahul C
Abstract: Image denoising is an important preprocessing technique in medical image analysis. The presence of noise in images can lead to degradation in its quality. Image denoising is the process of removing noise from an image and is basically a tradeoff between noise removal and preservation of significant image details. This paper presents a new sparse processing based denoising algorithm, the MMV-ICA (Multi-Median Variance-Independent Component Analysis) denoising algorithm. The MMV-ICA algorithm has been implemented and applied to medical images and the results are analyzed. Various noises which affect medical images are also considered. The proposed denoising algorithm is based on sparse and redundant representations over learned dictionaries. The dictionary is trained using the corrupted image. Thereafter, the dictionary is adapted to achieve sparse signal representations. MMV-ICA algorithm presented in this paper makes use of a patch based dictionary creation method. The paper presents the results of MMV-ICA denoising technique, which are found to be in par with the existing sparse based denoising methods.
Keywords: Sparse processing; Dictionary learning; Image Denoising; Independent Component Analysis (ICA.
Contact less non-invasive method to identify abnormal tongue area using k-mean and problem identification in COVID 19 scenario
by Pallavi Pahadiya, Ritu Vijay, Kumod Gupta, Shivani Saxena, Ritu Tandon
Abstract: Due to the spread of COVID-19 all around the world there is a need of automatic system for primary tongue ulcer, cancerous cell detection since, everyone dont go to hospital due to the panic and fear of virus spread. These diseases if avoided may spread soon. So, in such situation there is global need of improvement in disease sensing through remote devices using non-invasive methods. Automatic tongue analysis supports the examiner to identify the problem which can be finally verified using invasive methods. In automated tongue analysis image quality, segmentation and area of affected region plays an important role for disease identification. This paper proposes mobile based image sensing and sending the image to the examiner, examiner if finds issue in image may guide the user to go for further treatment. For Segmentation of abnormal area k-mean clustering is used with varying its parameters.
Keywords: Tongue diagnosis system (TDS); Image Acquisition; Thresholding; Segmentation; k-mean clustering; mobile app.
Real-time electrocardiogram monitoring for heart diseases with secured internet of thing protocol
by Trupti G. Thite, Daulappa G. Bhalake
Abstract: Real-time effective ECG data collecting, transmitting, and monitoring system with feature extraction is a big challenge in biomedical signal processing. The electrocardiogram is a widely used testing system to measure and analyse coronary heart diseases, i.e., cardiovascular diseases (CVDs). Heart rate remote monitoring under the service provided by hospital equipment is the current need to improve technologies. IoT enabled medical device helps efficiently to achieve this. To design such systems energy-efficient communication protocol, data-transfer minimisation, assurance of delivery (security), heterogeneous natures of the environment are necessary considerations. This paper outlines a literature survey of three main important areas; real-time ECG monitoring using wearable sensors, feature extraction and classification method for real-time ECG monitoring, and secured IoT protocol for real-time ECG monitoring.
Keywords: ECG; cardiovascular disease; CVD; tele-monitoring; real-time; tele-health; security; internet-of-things.
A new hybrid method for left ventricular analysis in cardiac cine MRI
by Sarra DALI YOUCEF, Messadi Mahammed
Abstract: The increase of deaths by cardiac arrest each year makes the computer-aided diagnosis a necessity for prognosis and treatment of cardiovascular diseases. The cardiac MRI is an imaging technique commonly used for the exploration of the heart. This technique has emerged as a reference for diagnosis of various cardiovascular diseases. In order to analyze the cardiac function in MRI image, the segmentation phase of the left ventricle (LV) is a necessary step to separate left ventricular region from the back ground. Wherefore, thresholding, region growing and active contour model are combined in our approach to obtain the left ventricle form exactly. Finally, the parameters such as end-diastolic volume (EDV), end-systolic volume (ESV) and ejection fraction (EF) are calculated for the LV function quantification. The whole process is applied to the Heart data base containing 18 patients where the experts manual contour is available. Our results show that our method gives an excellent segmentation of LV and a good correlation between our parameters and those obtained by the experts. We have found a correlation of 97% for EDV, 96% for ESV and 89% for EF. They confirm the accuracy of the proposed method and its eventual in aid of diagnosis.
Keywords: left ventricle; cardiac cine MR images; segmentation; thresholding; region growing; active contour; characterization.
Detecting Heart Ailments by Investigating ECG with Neural Networks
by Prabadevi B, Deepa N, KRITHIKA LB, RAVI R.A.J. GULATI, Sivakumar R
Abstract: Heart ailments or Cardiovascular Disease (CVD) are the diseases that incorporate the blood vessels or heart, which is common among various age groups. Though numerous techniques have been used to classify heart abnormalities, such as Classification and regression trees (CART) they are less accurate. Therefore, a technique for early detection of heart ailments with more accuracy is mandatory. A model has been designed and proposed to detect the heart ailments using three-layered neural networks for better accuracy. Electrocardiogram (ECG or EKG) is used to identify arrhythmia (irregular heartbeat) accurately, and the UC Irvine (UCI) arrhythmia dataset of ECG reports are used to implement a classification for different types of heart abnormalities.
Keywords: cardiovascular disease; electrocardiogram; Networks; Arrhythmia; Classification.
ANALYZING HOSPITAL FACTORS INFLUENCING INTERHOSPITAL SURGICAL SITE INFECTION RATES
by Vibha Gujar, Shankar Srinivasan, Dinesh Mital, Frederick Coffman
Abstract: The surgical site infection (SSI) prevention enactments in the hospitals require comprehensive infection surveillance and control. Each hospital system in the United States displays its own population demographics, pathogenic profiles and surgical volume based on varying geographical location. Remarkably, to stop current challenges like penalties due to poor quality in care, the methods to detect SSI effects and care quality require updates based on meaningful assessment of rates between the hospitals by adding more features. Therefore, utilizing the risk factors adjustment, this retrospective analysis aimed to analyze SSI patients and compare interhospital at-risk individuals based on hospital-related features. The descriptive and regression analysis for each hospital size (small:<250 beds, medium: up to 450 beds, and large: 450+ beds) demonstrated a significant influence of surgical categories, pathogen, hospital location and teaching status on the SSI rates. It distinctively identified patients with cardiovascular and respiratory surgical procedures more in rural than urban hospitals as the high-risk interhospital distinguishing clusters. Though infection rates conversed the population prevalence, adjusting the sociodemographic and other hospital characteristics for the case-mix, regression helped recognize the at-risk patients broadly. The findings from this study, thus, can help hospital organizations to define more case-mix features to device premature cautioning systems before discharge and better tracking approaches.
Keywords: Surgical site infection; post-surgical infection; catheterization; hospital size.
Cell-Seeded Small Intestinal Submucosa (SIS) as a Synthetic Vascular Graft for Implantation in Dogs
by Mohsen Ahmadi, Behnam Molavi, Ali Ghiaseddin, Shahram Rabbani, Hosein Ahmadi Tafti, Reza Ghiassi, Abtin Mamdouh, Shapoor Shirani
Abstract: We investigated the use of small intestinal submucosa (SIS) as a synthetic vascular graft for implantation in a canine model. Three months after implanting the graft in the thoracic aorta of the dogs, animals were sacrificed and grafts were removed for mechanical testing and cell differentiation analysis. The results showed differentiation of bone marrow cells into endothelial smooth cells and lower levels of vimentin, vascular endothelial growth factor receptor (VEGFR), cellularity, and collagen amount in the graft compared to the aorta. In terms of mechanical properties, the grafts were significantly more rigid than the natural aorta. Finally, while the coexistence of differentiated cell layers containing fibroblast, smooth muscle cells and endothelial cells made the SIS a potentially effective artificial graft in the dog model, lack of enough flexibility of the graft remains a problem relative to clinical use of the grafts.
Keywords: Intestinal submucosa; Biodegradable scaffold; Cell seeding; Synthetic vascular graft.
Performance Evaluation of Optimized SVM for Classification of Brain Tumor
by Arun Kumar, M.A. Ansari, Alaknanda Ashok
Abstract: In today's scenario, machine learning tools are most widely used for the classification of images in the field of medical science. Support Vector Machine(SVM) is one of them most popular and highly used for such classifications. Further, such classifications are highly related to the number of features selected from any medical image. The computation time and the memory required for the successful implementation of any classification tool is directly dependent on the number of features. So, in order to get the more accurate classification results, the features of the medical image must be optimized. The present study mainly aims on the development of an improved classification technique by combining with some optimization approaches. In this study, support vector machine is implemented for the classification of the brain tumor by optimizing the features of the Magnetic resonance imaging (MRI) images using three different optimization approaches namely, Particle swarm optimization, Grey wolf optimization and Firefly Algorithm. The results obtained from this study depict that support vector machine along-with the grey wolf optimization provides more accurate classification of the brain tumor with an accuracy of 96.8%
Keywords: Magnetic resonance imaging;Classification;Optimization; Brain tumor;Supprot Vector Machine.
Impact of COVID-19 on Individuals Mental Health and Preventive Health Behaviors: A Conceptual Framework
by Rajesh Pai, Naganna Chetty, Sreejith Alathur
Abstract: The Corona Virus Disease (COVID-19) is a pandemic that facilitate a confrontation space for scientific and social existence of human frontiers. The rapid spread and mortality rate of COVID-19 and the preventive measures including social distancing and its impact on economy, developed an unprecedented consciousness around the globe. It has created an effect on the mental health of individuals employed across various sectors and is outlined in this study. There is currently an inadequate theoretical model that focuses on the comprehensive understanding of the psychology of preventive behavior during the outbreak of pandemics. In this study, a transnational model is delineated for assessing the adoption of preventive behavioral practices associated with COVID-19 pandemic. It uses the components derived from the theories of situational awareness and health belief model and literatures related to impact of containment strategies on various sectors. The contribution include policy recommendations that can be helpful for the healthcare professionals and government to control the disease spread.
Keywords: COVID-19; Health Belief Model; Situational Awareness; Mental Health; Preventive Health Practices.
Comparative analysis of various supervised machine learning algorithms for the early prediction of type-II diabetes mellitus
by Shahid Mohammad Ganie, Majid Bashir Malik
Abstract: Diabetes is one among the top 10 causes of death. Diabetes mellitus is a fatal disease that poses a unique and significant threat to millions of people over the globe. Despite the absolute truth about the statistical data of diabetes from various sources like the World Health Organization, International Diabetes Federation, American Diabetes Association, etc. there is a positive message that early prediction along with appropriate care, diabetes mellitus can be managed and its complications can also be prevented. Nowadays in healthcare sector, machine learning techniques are gaining immense importance through their analytical classification capabilities. Machine learning paradigms are being exploited by researchers for better prediction of diabetes to save human lives. In this paper, a comparison of different supervised machine learning classifiers based on the performance evaluation of various metrics for the early prediction of type-II diabetes mellitus (T2DM) has been performed. The experimental work has been successfully carried out using six machine learning classification algorithms. Among all classifiers, random forest (RF) performs better for predicting T2DM with an accuracy rate of 93.75%. In addition, ten-fold cross-validation method has been applied to remove the class biasness in the dataset.
Keywords: type-II diabetes mellitus; T2DM; machine learning; framework; logistic regression; LR; Naïve Bayes; NB; support vector machine; SVM; decision tree; DT; random forest; RF; artificial neural network; ANN.
Computational fluid dynamic analysis of carotid artery with different plaque shapes
by Raman Yadav, Sharda Vashisth, Ranjit Verma
Abstract: Plaque formation in the carotid artery results in carotid artery disease. Atherosclerotic plaque is mostly found at the branching and bifurcation of the artery. The present work investigates the effect of Wall Shear Stress (WSS) and blood flow through carotid artery under various stenosis shapes. Five plaque shapes are considered i.e. plaque at branching, plaque at bifurcation, cosine plaque, irregular plaque, blood clot in external artery. WSS and velocity of blood through stenosed artery was simulated and analyzed using ANSYS Fluent Computational Fluid Dynamics (CFD). Comparison of the wall shear stress at wall artery showed that the artery having blood clot has maximum WSS followed by plaque at bifurcation, cosine shape, irregular shaped stenosis and plaque at branching. It is found that shapes of stenosis play key role in WSS. As stenosis increases in artery WSS also increases. The velocity of flow across stenosis is highest for artery having plaque at bifurcation followed by plaque at branching, irregular plaque, cosine plaque and artery having clot.
Keywords: Wall Shear Stress; Atherosclerosis; Bifurcation; Stenosis; Plaque shape; Computational Fluid Dynamics.
Segmentation of Retinal Blood Vessel structure using Birnbaum-Saunders (Fatigue Life) Probability Distribution Function
by K. Susheel Kumar, Nagendra Pratap Singh
Abstract: Segmentation of the retinal vessel in an eye is a significant task. Retinal blood vessels contain essential information useful in the computer-based diagnosis of various retinal pathologies, such as diabetes, hypertension, etc. In this paper, a novel approach of Probability Distribution Function of Birnbaum Saunders (Fatigue Life) based on matched filtering methods and imported to improve the segmentation of retinal blood vessels concerning existing matched filter methods. In this paper, the retinal blood vessel segmentation divided into preprocessing matched filter-based proposed method and postprocessing. In the preprocessing stage, improve the retinal image quality a different process is known as Principal Component Analysis (PCA) is used to convert to grayscale, followed by a Contrast Limited Adaptive Histogram Equalization known as CLAHE to enhance the grayscale retinal image. For designing of the Birnbaum Saunders (Fatigue Life) based matched filter, suitable values of the different parameters are chosen based on a complete experimental analysis In postprocessing based on an optimization technique based on entropy and length filter for removing the outer artifacts. The proposed approach tested on retinal images of DRIVE database to measure the performance in term of Average True Positive Rate (ATPR), Average False Positive Rate (AFPR), Average Accuracy, Average Root mean square deviation (RMSD), Avg F1-Score and Receiver operating characteristic (ROC) curve plotted. Average Area under the curve (AUC) calculated. The results of values are obtained ATPR 71.39 %, AFPR 2.67 %, Average Accuracy 94.61 %, Average RMSD 0.0054, Average F1-Score 0.684 and Average AUC for DRIVE Dataset 0.9361 respectively.
Keywords: Birnbaum-Saunders (Fatigue) Probability Distribution Function; Matched filter; Retinal blood vessel segmentation; Optimal thresholding-based entropy.
Comparison of image reconstruction algorithms for finding impurities utilising EIT for clinical application in breast cancer
by Priya Hankare, Alice N. Cheeran, Prashant Bhopale
Abstract: Breast cancer is a common and life threatening disease if not treated in its early stage. Electrical impedance tomography is an imaging technique employed in medical field for analysis and diagnosis purpose for early breast cancer disease detection, which is based on voltage and current or impedance measurements. In this paper, 2-dimensional electrical impedance tomography database is used to study and implement various image reconstruction algorithms. The electrical impedance and diffused optical reconstruction software (EIDORS) of MATLAB toolbox is used to reconstruct images of circular phantom approximating a breast hypothetical model.
Keywords: electrical impedance tomography; EIT; tumour; phantom; image reconstruction.
Autism Spectrum Disorder Diagnosis and Machine Learning: A review
by Chandan Jyoti Kumar, Priti Rekha Das, Anil Hazarika
Abstract: Autism spectrum disorder (ASD) with global prevalence estimate of approximately 1%, makes it a major social health concern. To make the diagnostic process of ASD faster, convenient and more accurate the researchers have started to apply a dozen of machine learning techniques. This review considers major publications of last decade to identify various aspects of machine learning research in ASD diagnosis. Findings of diagnostic tools and techniques are highlighted so as to detect significant features for machine learning models. Based on types of data, the article categorizes the diagnostic research in two broad categories: behavioral and neuroimaging. In addition, it explores the various findings of these behavioral and neuroimaging techniques in ASD subjects and makes a detailed analysis of performance of these techniques in combination with different machine learning models for ASD diagnosis. This article highlights key research fields of ASD and discusses potential research direction in the future.
Keywords: Autism Spectrum Disorder; Machine Learning; Neuroimaging; ASD Datasets.
A novel method to study resting-state and functional connectivity in infants using coherence analysis of EEG
by Hemang Shrivastava
Abstract: In this study, our goal was to study functional connectivity in infants using event-related potentials (ERPs) of electroencephalography (EEG). We hypothesised that coherence analysis of the power spectral density of tactile stimuli responses would differentiate preterm from full-term infants. In our knowledge, this is the first study demonstrating differences between resting state and tactile functional connectivity using touch stimuli, in preterm infants. We concluded that tactile brain connectivity in full-term infants is more efficient than preterm infants. No statistically significant differences were found in resting-state connectivity for full-term and preterm infants.
Keywords: functional connectivity; resting-state connectivity; coherence analysis; electroencephalography; EEG; event-related potential; ERP; infant brain development; somatosensory; connectivity networks; small world networks.
Curtailing Insomnia in Non-Intrusive hardware less Approach with Machine Learning
by K.V. Shriram, Sini Raj Pulari, Ragu Raman
Abstract: The significant challenges nowadays with the expanded utilization of the cell phone are restlessness and risk to mental health. Rest time is implied for the cerebrum to revive. If the rest designs are disturbed because of a nonstop outer aggravation, it upsets the profound rest. Most of us prefer music as the option to induce sleep and relax. Headphones or earphones are used for the same. It is shrewd to turn off the music after the individual rests, which majority of us don\'t do as we are as of now rested by at that point. This causes damage. Excessive usage of earphones or headphones is one part of it and unnecessary feed to the ears while sleeping shall trigger Noise-Induced Hearing Loss. Here, we propose a framework built with Machine Learning as key. This will guarantee that the music player is halted once the individual using has dozed. This ensures proper rest and forestalls sleep deprivation/ NIHL.
Keywords: Machine Learning; Insomnia; Sleep loss; Noise-Induced Hearing Loss; Technology for sleep; Hearing Loss;.
An affordable, intelligent, and fully functional Smart Ventilator System.
by Bharath Krishnan, Achuth Karakkat, Rohit Mohan Menon, K.V. Shriram
Abstract: Because of the Corona Virus Disease (COVID-19) pandemic scenario that the world is going through right now, there has been a surge in the requirement for emergency life support systems like ventilators. Conventional ventilators used in Intensive Care Unit (ICU)s tend to be bulky and expensive and demand high power consumption and trained experts to operate. The aim of the project is to deliver a solution for the growing demand for portable ventilators and a viable replacement for nurse assisted artificial resuscitation. Mechanical ventilation is the process of supplying scheduled breaths to a patient who lacks the ability to do the Work of Breathing (WOB) himself/herself. The pattern of breathing for every patient is identified using sensor(s) and the required volume of air is supplied by compressing a Bag Valve Mask (BVM) device. A machine learning algorithm learns the pattern of breathing and adjusts the pressure and volume controls specific to every patient. All operations and control mode switching for the device can be done using an Android app, hence making it user friendly.
Keywords: Covid 19; CoronaVirus; ICU; Ventilator; Smart Ventilator; Breathing issues;.
IHDPM: An Integrated Heart Disease Prediction Model for Heart Disease Prediction
by Abhilash Pati, Manoranjan Parhi, Binod Kumar Pattanayak
Abstract: The prediction of Heart Disease (HD) helps the physicians in taking accurate decisions towards improvement of the patient's health. Hence, Machine Learning (ML), Data Mining (DM), and Classification techniques play a vital role in understanding and reducing the symptoms related to heart diseases. In this paper, an Integrated Heart Disease Prediction Model (IHDPM) has been introduced for heart disease prediction by considering Principal Component Analysis (PCA) for dimensionality reduction, Sequential Feature Selection (SFS) for Feature Selection, and Random Forest (RF) classifier for the purpose of classifications. Some experiments are performed by considering the evaluative measures namely, Accuracy, Error rate, Precision, Recall or Sensitivity, F-Measure, True positive rate, True negative rate or Specificity, False positive rate, False negative rate, and Matthew's Correlation Coefficient (MCC) on Cleveland Heart Disease Dataset (CHDD) sourced from UCI-ML repository and Python language thereby concluding that the proposed model outperforms in comparison with other six conventional classification techniques. The proposed model will help out the physicians in conducting diagnosis of the heart patients proficiently and may be helpful in further investigations using different datasets related to heart disease and at the same time, it can be applicable in predictions of other chronic diseases like Diabetes, Cancers etc.
Keywords: Machine Learning (ML); Data Mining (DM); Classification Techniques; Sequential Feature Selection (SFS); Principal Component Analysis(PCA); Heart Disease Prediction.
Harnessing the power of machine learning for breast anomaly prediction using thermograms
by Aayesha Hakim, R.N. Awale
Abstract: Breast cancer is the most fatal cancer among women globally. Thermography provides an early sign of a developing abnormality based on the temperature changes in breasts. In this work, statistical features extracted from the segmented breast region are used for breast cancer prognosis. Machine learning algorithms like support vector machine (SVM), k-nearest neighbourhood (kNN), naive Bayes and logistic regression without and with principal component analysis (PCA) as a pre-cursor are applied to the extracted data to classify thermograms as malignant or benign. Classification was also performed using tree-based classifiers, namely, decision tree and random forest. This work indicates that thermal imaging is capable of predicting breast pathologies coupled with machine learning algorithms. The PCA-SVM model has the best accuracy, sensitivity, specificity and AUROC of 92.74%, 77.77%, 95.83% and 0.8699 respectively. Among tree-based classifiers, random forest classifier has the best accuracy, sensitivity, specificity and AUROC of 94.4%, 97.5%, 78.72% and 0.97961 respectively with five-fold cross validation. Our study produced competitive results when compared to other studies in the literature.
Keywords: breast cancer; breast thermography; infrared imaging; thermal imaging; machine learning.
Classification of ECG arrhythmia using significant wavelet-based input features
by Shivani Saxena, Ritu Vijay, Pallavi Pahadiya, Kumud Kumar Gupta
Abstract: This paper proposes an automated approach to classify ECG arrhythmia using wavelet transform and neural network. Wavelet-based optimal ECG feature sets are prepared followed by regression plots in curve fitting. These feature sets are further used for pattern recognition to distinguish in between normal or abnormal arrhythmia classes using multi-layer perceptron neural network (MLP NN). To evaluate performances of the designed classifier accuracy, selectivity and sensitivity parameters are measured. The average accuracy of the classifier is 99.05% which is comparatively higher than the existing methods with dependence on less input features.
Keywords: ECG arrhythmia; MLP NN; performance indices; regression plot; wavelet transform.
Managing infectious and inflammatory complications in closed kidney injuries on the basis of fuzzy models
by Nikolay Korenevskiy, Seregin Stanislav Petrovich, Riad Taha Al-Kasasbeh, Ayman Ahmad Alqaralleh, Gennadij Vjacheslavovich Siplivyj, Mahdi Salman Alshamasin, Sofia Nikolaevna Rodionova, Ivan Mikhailovich Kholimenko, Maxim Yurievich Ilyash
Abstract: The aim of the work is to reduce the time and cost of forecasting, preventing and treating patients with infectious complications with kidney injuries. The studies are based on the results of a retrospective analysis of medical records of 123 patients with various forms of kidney injury. Using the methods of analysis, it was shown that in order to achieve acceptable for practice quality prediction of possible infectious and inflammatory complications. The level of psycho-emotional stress and energy of biologically active points connected to the kidneys is used as additional prognostic signs, confidence in the correct prediction increases to 0.93.
Keywords: class membership functions; exploratory analysis; fuzzy logic; level of psycho-emotional stress; prognosis; risk of developing infectious and inflammatory complications; treatment regimens.
Analysis of body constitutions discrimination based on radial pulse wave by SVM
by Nan Li
Abstract: In this work, a new method for distinguishing humans physical constitution based on pulse information is proposed. Firstly, pulse data were collected, processed and pulse cycles were segmented. Secondly, time domain features, features coefficients, power spectrums and energy values of pulse wave were extracted and analysed, respectively.
Finally, pulse features were evaluated and classified to distinguish different body constitutions by SVM classifier. The experiment indicated that the features selected could be appropriately used to analyze the physical constitutions and can serve as the basis for research on constitution assessment based on Traditional Chinese Medicine pulse diagnosis.
Keywords: Traditional Chinese Medicine; Pulse Characteristics; Body Constitution; SVM.
Layer based Deep Net models for Automated Classification of Pulmonary Tuberculosis from Chest Radiographs
by Sushil Ghildiyal, Saibal Manna, Ruban N
Abstract: Tuberculosis (TB) is a highly infectious bacterial disease. However, it can affect any body part, but is majorly a lung infection; which is potentially fatal and contagious. Many of the deceased could have been saved, provided the infection would have been diagnosed at an earlier stage. Like most of the serious health issues, the recovery rate of a symptomatic TB patient completely depends on the early detection and treatment for the infection. Therefore, replacing a current computer aided diagnosis (CAD) systems, which are mostly based on manual features with CAD systems having automatic feature extractors based on deep learning, can provide aid in early detection and eventually in a curtailment of the disease. In this regard, a method to detect infection of tuberculosis, which uses deep learning, is presented in this paper. The said method uses Deep Neural Network to classify CXR images as normal or abnormal. CNN (Convolutional Neural Network), VGG16 (Visual Geometry Group) and HRNet (High Resolution Network) models are used and their performance has been compared based on the validation loss and validation accuracy. The HRNet provides 89.7% accuracy with comparatively less loss among the proposed algorithms. The models are also deployed in android application for active clinical trials.
Keywords: Tuberculosis; Deep Neural Network; CNN; VGG16; HRNet.
Outbreak Trends of Fatality Rate into Coronavirus Disease - 2019 using Deep Learning
by Robin Singh Bhadoria, Yash Gupta, Ivan Perl
Abstract: The World Health Organization (WHO) has declared the novel coronavirus as global pandemic on 11th March 2020. It is supposed to known from Wuhan, China and its spread is unstoppable due to no proper medication and vaccine. The developed forecasting models to predict the number of cases and its fatality rate for coronavirus disease 2019 (COVID-19) which is highly impulsive. This paper provides an intrinsic algorithms namely Linear Regression and Long short-term memory (LSTM) using deep learning for time series based prediction. It also uses the ReLU activation function and Adam optimizer. This paper also reports a comparative study on existing models for COVID-19 cases from different continents in the world. It also provide an extensive model that a brief the prediction about the number of cases and time for recovered, active and deaths rate till January 2021.
Keywords: Pandemic Analysis; Coronavirus Disease – 2019 (COVID-19); Linear Regression; Time Series forecasting; Long Short-Term Memory (LSTM); Deep learning;.
A Hybrid Random Forest based Feature selection model using Mutual Information and F-score for Preterm birth classification.
by Himani Deshpande, Leena Ragha
Abstract: Every womans body is unique and will have some features playing a vital role contributing towards a healthy pregnancy and manually it is difficult to decide the important features to be observed to prevent pregnancy complications. In this proposal, we have considered 21 physical features of 903 women of varied age groups, economic status and health conditions. Variation and Information based Random Forest(VIBRF) hybrid model using mutual information and F-score is applied to evaluate each feature looking into the variation within the feature and mutual information across the features. We experimented using various classifiers, and it is observed that Gaussian NB has shown the most significant improvement in terms of prediction accuracy, from 31% with all features to 80% with our feature selection process. Though SVM prediction accuracy is 84% it is observed AUC drastically improved for GNB by 10%. As it is a medical application, it is important to achieve higher AUC and so through this experiment, it is concluded that GNB performs better with the proposed model.
Keywords: Features Selection; F-score; Decision Tree; Random Forest; Hybrid Model; Preterm birth; Classification.
Perception and confusion of speech in Algerian school children wearing hearing aids
by Kamel Ferrat, Samir Benyahia
Abstract: The paper discusses the performance of hearing impaired Algerian children in perception of features of the Arabic phonemes in comparison to their counterparts with normal hearing ability. The acoustic analysis of collected data demonstrates a presence of several articulation disorders at primary school children with average deafness and wearers of hearing aids, such as: the substitution, elision, assimilation, addition and deformation of Arabic consonants in continuous speech. The consonants prone to these disorders are the back consonants mainly the fricatives and sibilants. Therefore, school rehabilitation services should take into account these constraints to achieve better schooling of hearing impaired schoolchildren.
Keywords: average deafness; acoustic analysis; hearing aids; primary school; Arabic language.
Design and Development of IoT - WBAN based Biomedical Solutions via Three - Tier Approach
by Sonal , S.R.N. Reddy, Dinesh Kumar
Abstract: This paper discusses a multi-sensor network based on the IoT-WBAN architecture, designed and developed using a threetier approach to simplify the configuration and networking of sensor nodes. The proposed framework has been developed to take
into account various real-time criteria, such as affordable, unobtrusive, non-invasive monitoring at anytime and anywhere. The device is made up of multiple wireless sensor nodes, each recording the various physiological parameters of the patient. At the
hub/aggregating unit at TIER-1, the separate data obtained are aggregated and then sent to the base station (TIER-2) for remote transmission (TIER-3). The base station functions as an intermediary point for the transmission of long-range data over the internet or cell network. In order to address the current constraints, numerous design challenges have been considered.
Keywords: Sensor Node Designing; IoT; WBAN; CHD.
Class prediction of the prevalent transmission mode of COVID-19 within a geographic area
by Donald Douglas Atsa'am, Ruth Wario
Abstract: This research developed a multinomial classification model that predicts the prevalent mode of transmission of the coronavirus from person to person within a geographic area, using data from the World Health Organization (WHO). The WHO defines four transmission modes of the coronavirus disease 2019 (COVID-19); namely, community transmission, pending (unknown), sporadic cases, and clusters of cases. The logistic regression was deployed on the COVID-19 dataset to construct a multinomial model that can predict the prevalent transmission mode of coronavirus within a geographic area. The k-fold cross validation was employed to test predictive accuracy of the model, which yielded 73% accuracy. This model can be
adopted by local authorities such as regional, state, local government, and cities, to predict the prevalent transmission mode of the virus within their territories. The outcome of the prediction will determine the appropriate strategies to put in place or re-enforced to curtail further transmission.
Keywords: COVID-19; transmission mode; multi-class prediction; predictive model; community transmission.
A Comprehensive Review on the Diagnosis and Testing Strategies for Coronavirus Disease (COVID-19).
by Ragul V, Vishnu Priya Veeraraghavan, KRISHNA MOHAN SURAPANENI, Shanmugarathinam A, Niyas Ahamed
Abstract: The COVID-19 outbreak has fashioned to severe threat to each and every individual in social and economic aspects in the country. This can be ascribed to the unreliable properties of COVID-19: it poses a unique standard of broadcast and death ratios. From this review, the probable property of these deadly transmissible viruses is related to that of SARS-CoV-2 as a fright zone of viruses. These agents can be effective and accurate identification which is target separation and management in the affected individuals are essential at the initial phase of viral attack. This is a very crucial and wide factor to know how it is different and dominant, to determine effective vaccines to avoid the transmission of these deadly causative agents. As of September 2020, more than 100 diagnostic kits and developing technologies were used for the detection of COVID-19 are surveyed in this review. The effective management and control of PZV and SARS-CoV-2 are more important to reduce the pandemic situation. The specified and accurate diagnostic and sensitivity materials are required to detect the above wide threat virus in the society by using nucleic acid based diagnosis which help to decrease the negative results. At last the drastic effect of this virus required dynamic defense in the entire world, The society entirely dependent on both private NGOs and government sections for the development of cost effective and constant testing kits in general PZV in future.
Keywords: COVID-19; SARS-Co-2; Panic Zone of Virus; Pandemic.
Automatic detection of Novel Corona virus (SARS-CoV-2) infection in computed tomography scan based on local adaptive thresholding and kernel-support vectors
by Ritam Sharma, Jankiballabh Sharma, Ranjan Maheshwari
Abstract: The recent studies have indicated the requisite of computed tomography scan analysis by radiologists extensively to find out the suspected patients of SARS-CoV-2 (COVID-19). The existing deep learning methods distribute one or more of the subsequent bottlenecks. Therefore, a straight forward method for detecting COVID-19 infection using real-world computed tomography scans is presented. The detection process consists of image processing techniques such as segmentation of lung parenchyma and extraction of effective texture features. The kernel-based support vector machine is employed over feature vectors for classification. The performance parameters of the proposed method are calculated and compared with the existing methodology on the same dataset. The classification results are found outperforming and the method is less probabilistic which can be further exploited for developing more realistic detection system.
Keywords: COVID-19; CT; artificial Intelligence; textural feature; adaptive thresholding; support vector machine.
Early Detection of Parkinsons Disease by using Neuroimaging and biomarkers through Hard and Soft classifiers
by Gunjan Pahuja, Bhanu Prasad
Abstract: Early and accurate detection of Parkinson's disease (PD) remains a challenge. Two prevalent approaches used for the detection of PD are: (i) Dopaminergic imaging using Single Photon Emission Computed Tomography (SPECT) with 123I-Ioflupane and (ii) Cerebrospinal Fluid (CSF) biomarkers. Striatal Binding Ratio (SBR) values are computed from SPECT and, in this research, it is found that if these SBR values are complemented with CSF biomarkers then these SBR values help increase the accuracy of early PD detection. In this study, SBR values for each of the four striatal regions are complemented with some CSF biomarkers to develop a model for the classification and prediction of early PD. A hard classifier is used for developing the classification submodel, and a soft classifier is used for developing the prediction submodel. The results indicate the effectiveness of the developed model.
Keywords: Parkinson’s Disease (PD); Striatal Binding Ratio (SBR); hard classifier; soft classifier; Multivariate Logistic Regression (MLR); risk prediction; biomarkers.
Computer Vision-Based Approach for Detecting Arm-flapping as Autism Suspect Behavior
by Esraa Tarek Ahmed Hassan Sadek, Noha Aly AbdElSabour Seada, Said Ghoniemy
Abstract: Autism spectrum disorder (ASD) is a neurodevelopmental condition that is characterized by frequent and restrictive motor activities, besides other social and communicative disorders. It is considered one of the most rapidly evolving neurodevelopmental disorders in children during the past few years. Repetitive motor behaviors, like arm-flapping and head rocking, may lead to attention distraction and self-injury in severe cases. In this research, a computer-vision-based neural network framework is proposed to automatically detect significant arm-flapping behavior in autistic children. The proposed framework goes through four main phases which are data pre-processing, pose estimation and skeleton representation, data post-processing, and action classification. The proposed framework was successfully tested on three different datasets and proved its applicability in real-world applications. The attained accuracy was better compared to that of the state-of-the-art methods. The proposed solution can be used to assist clinicians, parents, and caregivers to automatically detect this behavior to offer the child the appropriate medical care once a behavioral abnormality is detected.
Keywords: Autism spectrum disorder (ASD); Arm Flapping; Computer Vision; Neural Networks.
PATIENT RELATED RISKS AND OUTCOME ASSESSMENTS ON INTERHOSPITAL SURGICAL SITE INFECTION RATES
by Vibha Gujar, Shankar Srinivasan, Dinesh Mital, Frederick Coffman
Abstract: This paper investigates the effect of hospital size (sizes: small:<250 beds, medium: up to 450 beds, and large: 450+ beds) and surgical site infection (SSI) on patient-related risks and treatment outcomes using multi-level hierarchy method. We employed the National Inpatient Sample (NIS) data, available through the Healthcare Cost and Utilization Project (HCUP-NIS, the year 2008 2012) as secondary data with an updated set of case definitions. The SSI risks and outcomes were assessed using regression and propensity score matching analysis. Out of the total 222,845 SSI patients, unadjusted prevalence rates (per 100 procedures) were 2.9% in small, 2.62% in medium, and 2.65% in large hospitals (p<0.0001). Small hospitals at rural locations charged SSI patients up to 33.4% (average: $30,891), about 12% more than urban hospitals within the first 15-30 days of hospital stay. However, results indicate that large hospitals were significantly efficient in infection prevention and had a lesser cost burden than small hospitals. The association of SSI and its impact on outcomes were relatively uniform, but the consequence appeared to be mediated by age, disease severity, neighborhood or economic conditions, elective, and transferred admissions for specific hospital size. Surveillance and policy targeting at refining the efficiency of infection prevention should thus focus on smaller hospitals.
Keywords: Surgical site infection; hospital size; secondary data; propensity score; case definition; total charges.
Application of chaos theory for arrhythmia detection in pathological databases
by VARUN GUPTA
Abstract: To handle the current pathological situation of heart-related diseases, various techniques belonging to automatic Electro-Cardio-Gram (ECG) signal analysis are already available but have not succeeded. In this paper, Savitzky-Golay Filtering (SGF) and Support Vector Machine (SVM) techniques are used for preprocessing and classification purposes. Feature extraction algorithms play a vital role in biomedical signal processing (BSP). For that purpose, the chaos analysis theory is used as a feature extraction tool on different pathological datasets obtained from different cardiology labs to classify different arrhythmia types. The effectiveness of the proposed methodology is evaluated on different performance evaluating parameters viz. sensitivity (Se), accuracy (Acc), and duplicity (D). The proposed methodology presented Se of 99.87%, Acc of 99.72%, and D of 0.066%.
Keywords: Electro-Cardio-Gram (ECG) signal; Heart-related diseases; Biomedical signal processing (BSP); Chaos analysis.
Adaptive Neuro-Fuzzy Inference System for the diagnosis of non-mechanical low back pain
by Mehrdad Farzandipour, Ehsan Nabovati, Esmaeil Fakharian, Hossein Akbari, Soheila Saeedi
Abstract: Back pain is one of the most important causes of disability. Clinical Decision Support Systems (CDSSs) can help physicians diagnose diseases with greater precision. This study designs and implements a CDSS to diagnose non-mechanical Low Back Pain (LBP), including spinal brucellosis, ankylosing spondylitis, spinal tuberculosis, and spinal osteoarthritis using an Adaptive Neuro-Fuzzy Inference System (ANFIS). The highest corrected classification percentage was related to Spinal brucellosis (82.8%), and CDSS was able to differentiate four non-mechanical LBP types.
Keywords: Clinical Decision Support System; Non-Mechanical Low Back Pain; Adaptive Neuro-Fuzzy Inference System; Diagnose.
ICG Signal Noise Cancellation Algorithms for Non-invasive Hemodynamic Monitoring
by Hadjer BENABDALLAH, Salim KERAI
Abstract: Impedance Cardiography (ICG) non-invasive technique is a new way in the medical field, used for diagnosis and monitoring of cardiovascular disorders. The signal processing domain developed several denoising techniques applied for the cancellation of respiratory and movements artifacts that corrupted the acquired signal shape. For this main purpose, our paper-based on a comparative study between different type of adaptive filters and Savitzky-Golay (SG) filtering, those are applied to the sensed impedance to noise removal for hemodynamic parameters monitoring. The results demonstrated in this work are more suitable; it confirmed that the proposed SG denoising filter outperformed the other techniques cited in state of art in terms of preserving information carried over on the signal when the mean error rate of the SG technique is equal to 0.00038 %.
Keywords: Impedance Cardiography; Adaptive Filters; Savitzky-Golay; Mean Error Rate.
Improving prediction of one-year mortality of acute myocardial infarction using machine learning techniques
by Mirza Touseef, Najla Raza, Adeel Zafar, Muhammad Zubair, Saad Zafar
Abstract: The purpose of our study was to improve the prediction of one-year mortality for patients with acute myocardial infarction (AMI). We implemented and compared four classical machine learning algorithms and one deep neural network algorithm. For evaluation metrics, we used accuracy, F1-measure, precision, recall, and area under the receiver operating characteristic curve (AUC). Random forest achieved the best performance based on an AUC of 0.98 with an accuracy of 92%. Results show that our model can predict one-year mortality of AMI with an improved AUC and accuracy using a minimum number of features as compared to previous related studies.
Keywords: machine learning; deep neural networks; acute myocardial infarction; AMI; mortality prediction; cardiovascular diseases; CVDs.
Estimation Techniques of Vocal Fold Disorder: A Survey
by Satyajit Pangaonkar, Reena Gunjan
Abstract: Voice disorder of the speech has been observed as an utmost critical issue in both children and elders. Complexity, limited experimentation and predictions by the physicians during the analysis are main causes for inaccurate diagnosis and wrong information about the dysfunction. The research studies are limited to elementary analysis and diagnosis level for the physicians & therapists due the restrictions of skilled specialists and the expenses involved. This paper deals with the evaluation and facts to understand software tools, different acoustic, non-acoustic and non-invasive approach. This can be observed in terms of sensors and image extraction analysis to deal with the pathologies and to overcome the problems which will be beneficial to the speech language pathologists, clinicians, therapist and vocalist.
Keywords: FonaDyn; PRAAT; MDVP; TERC; PVDF; P-MIC; HSV; VKG.
A new Parkinson detection system based on evolutionary fast learning networks and voice measurements
by Ayoub Bouslah, Nora Taleb
Abstract: Parkinsons disease (PD) is become the second most neurological syndrome of the central nervous system after Alzheimers disease. It causes diverse symptoms which include bradykinesia (slowness of movement), voice impairments, rigidity, tremor, and poor balance. The PD recognition system based on voice is founded a noninvasive alternative, but involves rather complex measurements or variables. Therefore an attention is required toward new approaches for better forecasting accuracy. In this paper, an optimal Fast Learning Network (FLN) based on genetic algorithm (GA) was established as PD Diagnosis system. FLN is a double parallel feed-forward neural network structure, as a matter of fact, GA for features reduction and hyperparameters optimization of the FLN which is used as a predictive model. Finally, the conducted experiments on the Parkinson data of voice recordings over Ten Fold Cross-Validation show that proposed system is less complex and also achieved better average classification results with an accuracy of 97.47 %. At the same time, its effective in automatic identification of important vocal features. Moreover, the highest average degree of improved accuracy was (2.1 %) compared with other familiar wrappers including Support Vector Machine and K-nearest Neigbors in the similar conditions.
Keywords: Parkinson’s disease (PD); Fast Learning Network(FLN); genetic
algorithm (GA); features reduction; hyperparameters optimization; predictive
model; PD Diagnosis system.
Classification of Electroencephalography Signals using Three-Dimensions Convolution Neural Network with Long Short Term Memory Architecture
by Viet Quoc Huynh, Hoang-Thuy-Tien Vo, Thu Anh Nguyen, Tuan Van Huynh
Abstract: This research attempts to use the electroencephalography signals based on physiological signals extracted from the database for emotion analysis to classify the emotion of the subjects by classifier neural network algorithms. In this work, two types of neural network including 3D convolution neural network and hybrid network (3D convolution neural network model combined with long short term memory architecture) were applied to train and test its ability of emotion states classification. As a result, the hybrid network gave the most efficient classification with an accuracy of around 80\%, which was better than other algorithms such as Support Vector Machine, Random Forest, Convolution Neural Network. Furthermore, the results also showed that the accuracy achieved differently at various frequency bands, in which Delta frequency band gave the highest accuracy. Combining signals of different frequencies helped to improve the classification efficiency.
Keywords: Electroencephalography; Convolution neural network; LSTM; Hybrid network.
Current status of Hydroxychloroquine and Azithromycin for the Treatment of COVID-19: An Observational Review
by PRATEEK PRATYASHA, ADITYA PRASAD PADHY
Abstract: Though effect coronavirus has known to be a catastrophic pandemic since 100 of years, Severe Acute Respiratory Syndrome-2 coronavirus (SARS2-CoV) was first claimed to be emerged in December 2019 at the city of Wuhan, China. Abruptly, the virus dominated more than 218 countries with 157,566,607 confirmed cases and the death figure has reached nearly 32,84,551 till time. Recently the pandemic is getting worse day by day, people are suffering from hypoxia and server respiratory problem despite of the unstoppable service of healthcare sector. Prior concern behind this emergency is that, till date researchers and scientists were failed to invent any productive pharmaceutical treatment to weed out the infection completely. Although, vaccination is publically available; but it is applicable only for precaution purpose and not evident for preventive measures. This review focuses on therauptiuc status to control the severity of SAS2-CoV agent. The approach aims at implicating a low toxic metabolite anti-malarial drug, Hydroxychloroquine combined with an antibiotic called Azithromycin for the treatment of acute respiratory disturbance and hypoxia. This article briefly demonstrates the phramaco-potential of both these medications, their effects on patients based on a clinical observation and ongoing status of dosage to validate its implication.
Keywords: Severe Acute Respiratory Syndrome2-CoV (SARS2-CoV); COVID-19; Hydroxychloroquine; Azithromycin; Choloroquine.
Comparison between whole exome sequencing and the FoundationOne
by Catherine Wellan, Hillary Rayo, Shashi Mehta
Abstract: This systematic review and meta-analysis examines the question of how effective the FoundationOne
Keywords: tumor mutation burden; TMB; whole exome sequencing; WES; FoundationOne; targeted panels; next-generation sequencing; solid tumors; immune checkpoint blockade therapies; ICB; correlation; method comparison studies.
Identification of stroke using deepnet machine learning algorithm
by Abdulwahhab Alshammari, Noorah Atiyah, Hanoof Alaboodi, Riyad Alshammari
Abstract: Stroke is a disease that impacts individuals of all races, genders, and backgrounds. To combat the rising prevalence of the disease, the creation of accurate diagnostic tools is paramount. This paper uses two machine learning models, deepnet and decision tree, to assess the health record data from the Ministry of National Guard Health Affairs, Saudi Arabia. Deepnet outperformed the decision tree in accurately predicting stroke and stroke mimic. Deepnet achieved an accuracy of 92.35%, while decision tree achieved 90.8%. Future application of machine learning on stroke and stroke mimic diagnosis holds great potential in public health reform, patient empowerment, and minimising healthcare burden. This paper proposes building a national centralised semi-detection stroke data management framework to create a national platform in the diagnoses, acute, and long-term treatment of stroke.
Keywords: stroke; machine learning; identification.
Hepatitis C virus and cirrhosis: an analysis of incidence and cost of management of disease in the USA
by Pratima Tandon, Dinesh Mital, Shankar Srinivasan, Riddhi Vyas
Abstract: This study utilised the national (nationwide) impatient sample (NIS), which is part of the healthcare cost and utilisation project (HCUP). A study was conducted to determine the incidence and cost of management of Hepatitis C (HC) and cirrhosis (C) patients. The incidence rate of HC is 34% and C is 30%. HC is higher in the age group of 21-65-year old's whereas C prevalent in people that are 66-80+ years. The HC count is higher for Medicaid and, C is for Medicare payment methods. HC patients have a high Emergency admission rate. C patients show higher rate to transfer to a short-term hospital, home
healthcare, length of stay in the hospital, death rate during hospitalisation, frequency of biopsy procedures, charges for Liver Transplant, charges for admission to the trauma Centre as well as emergency admission. 62% of males have HC and 61% have C. For females 37% have HC and 39% C. HC and C is higher in Urban locations and low-income group. In average liver transplant is higher for the age group of 21 to 80 years. Asian people are less predisposed to getting HC.
Keywords: hepatitis; hepatitis c; cirrhosis; SPSS; data analysis; HCUP; NIS; liver disease; cost; charges; length of stay; health outcomes; incidence of hepatitis.
Clinical decision support system for the diagnosis of Hepatitis C virus
by Pratima Tandon, Dinesh P. Mital, Shankar Srinivasan, Riddhi Vyas
Abstract: A new clinical decision support system (CDSS) was developed using Exsys Corvid for the diagnosis of Hepatitis C virus. This CDSS is medically accurate and can guide healthcare professionals through the diagnostic process. Corvid Exsys rule-based system is used for building automated expert systems. The software utilises backward and forward chaining techniques. All the questions asked by the system during the diagnosis process are based on the clinical literature. The system can guide a clinician through the diagnostic process to achieve hepatitis results. Decision-making expert system was successfully developed to diagnose the Hepatitis C virus.
Keywords: hepatitis; Hepatitis C; clinical decision support system; CDSS; decision support systems; jaundice.
Identification and classification of schizophrenic speech using convolutional neural network for medical healthcare
by Akshita Abrol, Nisha Kapoor, Parveen Kumar Lehana
Abstract: Schizophrenia is a brain disorder that significantly affects the quality of life of affected individuals. One of its prominent symptoms is the induction of changes in the acoustics of the patients. In the absence of definite methods for its diagnosis, speech analysis can help in the preliminary screening of the patients. In this paper, an automated method using deep learning for differentiating between individuals with schizophrenia and psychosis from healthy individuals is suggested. Using convolutional neural networks with speech spectrograms as input, a classification accuracy of 87.01% has been obtained for levels of schizophrenia and 95.26% for differentiating between schizophrenic and healthy speech.
Keywords: schizophrenia; convolutional neural network; CNN; deep learning; spectrograms.
Automatic speech classification in school children with cleft (and lip) palate
by Khaled Baazi, Mhania Guerti
Abstract: The automatic detection of emphatic occlusive [t] in pathological speech (PS) in schoolchildren with cleft lip and palate can provide diagnostic information to clinicians and speech therapists. We propose an automatic classification system for PS by the decision tree method (DT) to use it in rehabilitation in Algerian hospitals. Acoustic analysis was applied in order to extract the relevant acoustic characteristics of this type of PS. The results showed that the DT adapts well to the classification of the PS with a significant rate ratio (%) of the PS, for the case of the long vowels 87% and 85% for short vowels.
Keywords: cleft palates; pathological speech; decision tree; classification; school children.
CIAE: class imbalance aware ensemble framework to predict drug side effects
by Kanica Sachdev, Manoj Kumar Gupta
Abstract: The binding of the drug compounds to certain biological off target proteins causes undesirable side effects or drug toxicology. The determination of drug toxicology at the early steps of drug development would help to economise on money as well as time. The paper proposes a novel framework, class imbalance aware ensemble (CIAE), for the identification of drug side effects using ensemble learning. It employs the related side effect information of the drugs to predict novel side effects. An eminent cause of the low performance of the machine learning based methods is the presence of class imbalance in the data. The proposed framework efficiently addressees this issue to improve the predictor performance. A comprehensive comparison of the method with the state of the art classifiers shows that the proposed framework yields better results for drug side effect determination.
Keywords: biological targets; class imbalance; drugs; drug side effects; drug toxicology; ensemble classifier; machine learning.
Special Issue on: Security and Privacy Concerns on Electronic Health Records
Integration of Intelligence in software development process for implementing a secure healthcare system - a review
by N. Asha, Siva Rama Krishnan, J. Gitanjali
Abstract: The implication of artificial intelligence (AI) has made substantial evolution in the modern environment. It has innovatively transformed the technical world and has assimilated everything in the real life. The impact of AI in the software development and process has significantly improved the progression with its cognitive behaviour. In modern-days, there is a remarkable increase in the data management; computational vitality has risen at reduced cost. Lot more breakthroughs are happening in technology, AI is performing all these seamlessly. To present the overall role of AI in software development and process, we have analysed the state-of-the-art of AI in software development for privacy in healthcare. The work is exemplified by recent advances in product development process and AI for securing medical data. Further, we also investigate on various phases of software development process where AI can be embedded for developing an efficient and secured healthcare system especially in the analysis of electronic medical records (EMR). We also discuss the challenges in incorporating AI in healthcare application.
Keywords: artificial intelligence; AI; security; privacy; healthcare; fault detection; electronic medical records; EMR.
Developing a biotechnical scheme using fuzzy logic model for classification of severity of pyelonephritis
by Nikolay Korenevskiy, Seregin Stanislav Petrovich, Riad Taha Al-Kasasbeh, Ayman Ahmad Alqaralleh, Sofia Nikolaevna Rodionova, Ashraf Adel Shaqadan, Ilyash Maxim Yurievich, Mahdi Salman Alshamasin
Abstract: The aim of the work is to develop fuzzy logic model to process health data involving oxidative indicators in patients with pyelonephritis to predict the severity level of pyelonephritis as severe and purulent forms. A 13 immunity and oxidative health indicators (lipid peroxidation) are used for classification of disease severity. A control sample of patient's is analysed to develop class's and experienced physicians are consulted to modify considered class limits. The fuzzy logic model gives high accuracy in diagnosis of serous and purulent pyelonephritis in patients with urolithiasis. Verification of the results of the operation of the obtained decision rules on the control sample showed that the proposed method's diagnostic efficiency reaches 93%, which is acceptable for use in medical practice.
Keywords: pyelonephritis; serious and purulent form; fuzzy logic; mathematical models; differential diagnosis; oxidative status.