International Journal of Medical Engineering and Informatics (39 papers in press)
Scaling-up Spatiotemporal Dynamics of HIV/AIDS Prevalence Rates in Sub-Saharan Africa Countries
by N. Kunene, William Ebomoyi, T.S. Gala
Abstract: With approximately, 2 in 3 global HIV/AIDS cases, Sub-Saharan Africa countries are enduring enormous HIV/AIDS disease burden. This study used Geographical Information System (GIS) to investigate the spatiotemporal of variability of HIV/AIDS prevalence rates of sub-Saharan African countries to provide hints of geographically resolved intervention strategies. Data derived from estimates based on national population and antenatal care (ANC) surveillance sites were acquired from the UNAIDS Global Report and World Reference Database and used for geospatial Analysis, longitudinal study and modelling. Accordingly, in sub-Saharan African, on average, 5.75% of the adult population were infected in 2014.
Keywords: HIV/AIDS; Health GIS; Geospatial Analysis; Trend Analysis; Sub-Saharan Africa.
Grove: An auxiliary device for sympathetic assessment via EDA measurement of neutral, stress, and anger emotions during simulated driving conditions
by Jonathan Shi Khai Ooi, Siti Anom Ahmad, Asnor Juraiza Ishak, Khairun Nisa' Minhad, Sawal Hamid Md Ali, Yu Zheng Chong
Abstract: Cognition, emotion, and mood are one of the most researched topics in psychophysiological signal study. Heart rate, skin conductance, and skin temperature are popular measures of understanding autonomic nervous systems. These measures are tightly related to sympathetic and parasympathetic nervous system, which regulates human emotion. Stress and anger affect driving task and contribute to the high number of road crashes. This study utilized electrodermal activity (EDA) to differentiate stress and anger from the neutral emotion of drivers while performing a simulated driving task. Twenty healthy subjects participated and the experiment protocol was approved by Ethics Committee for Research Involving Human Subjects, Universiti Putra Malaysia. Mean Power spectral density (PSD) of EDA signals were statistically compared between emotion groups with repeated-measures ANOVA and Bonferroni post hoc test. A significant difference (p <0.01) was observed between neutral-anger and neutral-stress groups, whereas no significant difference (p >0.01) was noted between stress-anger groups. Promising classification accuracy was achieved between emotion groups with support vector machine (SVM) classifier at 10-fold cross-validation.
Keywords: anger; cognition; driving; electrodermal activity (EDA); emotion; skin conductance response; stress.
ADVANCEMENTS IN EEG SOURCE LOCALIZATION METHODS
by Sanjeev Nara, Poonam Sheoran
Abstract: Electroencephalogram (EEG) signals represent the neuronal activity of brain. These signals are recorded by placement of multiple electrodes over the scalp or from cortex of the brain under the skull. These signals have important applications in biomedical and clinical field but most applications fail to take benefit of all the datas available from the information, particularly about the location of dynamic sources in the brain. Localization of sources of brain signal is important for the study of brain physiological, mental, pathological, and functional abnormalities, and problems related to various body disabilities. Their ultimate aim is to specify the sources of abnormalities such as tumors and epilepsy. This paper provides comprehensive overview of the different traditional and latest methods used for the EEG source localization.
Keywords: Electroencephalogram; Source localization; Brain abnormalities.
Segmentation of Coronary Artery Tree from Computed Tomography Angiography using Region Growing Method
by Marwa Shams, Mohammed A.-M. Salem, Safwat Hamad, Howida A. Shedeed
Abstract: Recently automated analysis of medical images becomes important for easier and faster clinical diagnosis. Identifying human organs is the key component for such analysis, i.e. segmentation of the anatomical structures from medical images. Coronary arteries segmentation gained wide interest in old and recent scientific research, thus various methods have been developed for segmenting coronaries from different cardiac imaging modalities. This paper provides a review of studies based on Region Growing (RG) method in segmentation of coronary arteries from computed tomography angiography (CTA). The main objective of this paper is to highlight the different perspectives of applying RG in the segmentation process. Firstly, medical background is provided to coronary disease, CTA and region growing algorithm explanation. Finally, the studies are compared to each other according to the selection of seed points, detection of seed points, preprocessing and enhancements, RG segmentation process and finally the post-processing.
Keywords: Region Growing; RG; Coronary Arteries; Vessel Segmentation;
Computed Tomography Angiography; CTA; Medical Imaging; Cardiology.
Analysis of Polycystic Kidney Disease in Medical Ultrasound Images
by Prema T. Akkasaligar, Sunanda Biradar
Abstract: The growth of kidney diseases has gradually increased in recent years. Ultrasound imaging provides the internal structure of the body to detect eventually diseases or abnormal tissues non-invasively. Segmentation of required region in ultrasound images is one of the challenging tasks. The proposed method focuses on classification of medical ultrasound images of kidney as cystic and polycystic types. Segmentation is performed using gradient vector force (GVF) snakes. Before segmentation, speckle noise is removed using Gaussian filter and contrast is enhanced. We have segmented normal, cystic and polycystic kidney ultrasound images effectively using GVF snakes. We have also carried out segmentation using morphological operations which requires an user intervention during the process of segmentation. Geometrical features are used with k-NN for classifing medical US images of kidney as normal, single cystic and polycystic types for segmented regions .The proposed method has applications in analysis of organ morphology and realizing quantitative measurements.
Keywords: GVF snakes; morphological operations; medical ultrasound image of kidney; polycystic kidney disease.
The Packet Wavelet Transform (PWT) in the analysis of Phonocardiograms (PCGs) Signals Aortic Stenosis (AS) and Mitral Stenosis (MS).
by Meziani Fadia, Debbal Sidi Mohammed El Amine
Abstract: Heart murmurs are the first signs of cardiac valve disorders. Indeed, heart auscultation has been recognized for a long time as an important tool for the diagnosis of heart disease; it is the most common and widely recommended method to screen for structural abnormalities of the cardiovascular system. Detecting relevant characteristics and forming a diagnosis based on the sounds heard through a stethoscope, however, is a skill that can take years to be acquired and refine. The efficiency and accuracy of diagnosis based on heart sound auscultation can be improved considerably by using digital signal processing techniques to analyze phonocardiographic (PCG) signals.rn The packet wavelet transforms (PWT) was successfully used as a feature to distinguish different heart sounds. Its introduced to analyze heart sounds and the feasibility of using in identification of three types of phonocardiogram's signals: three types of diseases were studied: Normal N, aortic stenosis AS and mitral stenosis MS. The Normal case present three cases, for the second and third each one presents six different cases. The importance of application of the packet wavelet transform (PWT) to analyze of the pathological severity of the aortic stenosis AS and mitral stenosis MS was presented. Then, the calculation of various parameters was performed for each patient. This study examines the possibility of using the packet wavelet transform (PWT) in the detected and analysis of Aortic Stenosis AS and Mitral Stenosis MS.rn
Keywords: Phonocardiogram; Aortic Stenosis; Mitral Stenosis; Murmur and Heart sounds; Discrete Wavelet Transform; Packet Wavelet Transform; Energy ,Ratio of Energy; Parameters; Measurements and Statistics..rn.
Watermarking Medical Images with Patient Identification to Verify Authenticity
by Sameh OUESLATI
Abstract: In medical imaging, it has been shown that watermarking can improve data protection and content enrichment. In this work we present an adaptive watermarking algorithm which exploits a wavelets-based human visual system (HVS) and a Fuzzy Inference System (FIS) to embed digital watermark while modifying frequency coefficients in discrete wavelet transform (DWT) domain. The main goal of the algorithm is to provide a more robust and imperceptible watermark. The application should be capable of handling the tradeoff between those two. This trade off is met when the position of embedding the watermark is optimally selected. In this paper, the FIS and HVS are combined to control and generate the watermark weighting function to embed the imperceptible watermark. Based on the experimental results, it is shown that the implemented watermarking algorithm is imperceptible and robust to some normal attacks such as JPEG Compression, Gaussian noise, Gaussian Blur, median filtering, and rotation.
Keywords: medical images; the human visual system; digital watermarking; fuzzy inference system; imperceptible; robust; weighting function.
An Adaptive Thresholding Technique for QRS-Complex Detection in ECG Signal Based on Empirical Wavelet Transform
by Trunal Jambholkar, Barjinder Singh Saini, Indu Saini
Abstract: ECG signal has been used as a diagnostic tool and its measurement become routine part of any complete medical evaluation. Since the QRS complex varies with different cardiac health conditions, therefore efficient and automatic detection of QRS complex and R-Peak is essential for reliable health monitoring. In this work an Empirical Wavelet Transform (EWT) based Algorithm has been used for accurate detection of QRS complex. EWT is one of the Adaptive time-frequency data analysis method. EWT is a fully data driven mechanism suitable for nonlinear and non-stationary data processes. In first step of this method decomposes the ECG signal into set of the AM-FM components called modes. First mode corresponds to the DC content of the signal decomposed and the later modes represent the different information content, contained in the signal decomposed. Therefore, this transform represents the original signal into different AM-FM components called Empirical modes. Later, Adaptive Thresholding is applied to its last mode to detection of QRS-complexes, which is nearly same as that of the original signal if we look at it visually. The proposed algorithm has been tested on the ECG signals of 48 records of MIT-BIH Arrhythmia database. Performance of proposed method has been measured on the basis of statistical parameters and gives the Positive Predictivity 99.82%, Sensitivity 99.93%, and Error Rate 0.24%. The proposed method is also tested on 20 Self-Recorded datasets with the sampling frequency of 1000Hz. The QRS-complex detection algorithm achieves 100% Sensitivity and 100% Positive Predictivity and zero Error Rate over the self-recorded.
Keywords: Empirical Wavelets; EWT; AM-FM components; QRS-Complex Detection; Adaptive Thresholding; ECG.
Adaptive filtering algorithm based on a wavelet packet tree for heart sound signal analysis
by Lotfi Hamza Cherif, S.M. Debbal
Abstract: In order to further highlight heart sound signals analysis, we developed an algorithm based on a wavelet packet tree; for possible discrimination depending on the severity of pathological cases for different heart sound signals. The algorithm functions select the most informative nodes combination of a wavelet packet tree as a basis for feature extraction. To generate this adaptive filter we need to compute the best sub tree of an initial wavelet packet tree with respect for entropy type yardstick, the node combination with the lowest total cost is selected. The decomposition into wavelet packet. Offers a wavelet library organized according to their time-frequency analysis and location properties and therefore of pass-band filtering, according to a binary tree architecture. This architecture makes it possible to implement algorithms for searching for adapted bases to both the desired time-frequency properties and the analyzed signal, which are conventionally called better bases.
Keywords: Phonocardiogram signal (PCG); Heart sounds; adapted bases;cost function; Adaptive filtering algorithm; Wavelet packet transform (PWT).
Cuckoo search based modified Bi-Histogram Equalization method to enhance the cancerous tissues in Mammography images
by Krishna Gopal Dhal, Mandira Sen, Sanjoy Das
Abstract: In this study novel variants of Histogram Equalization (HE) have been proposed by using proper histogram segmentation techniques and then incorporating weighting constraints to each sub-histogram independently to maintain the proper contrast. To segment the histogram properly; Otsu method, Kapurs Entropy and Gray Level Co-occurrence Matrix (GLCM) based entropy methods have been applied. Optimal weighting constraints have been computed by applying one existing modified Cuckoo Search (CS) algorithm. All variants are successfully applied to enhance the cancerous tissues of the mammogram images. Fractal dimension (FD), Entropy and Quality Index based on Local Variance (QILV) have been employed to measure the efficiency of all proposed methods. Experimental results prove the supremacy of the proposed methods over existing methods.
Keywords: Mammogram image Enhancement; Histogram Equalization; GLCM; Entropy; Fractal Dimension; Cuckoo Search.
Impact of Affective Picture and Music Stimuli on Autonomic Responses: Characterization of Pauses between Emotion Blocks
by Ataollah Abbasi, Ateke Goshvarpour, Atefeh Goshvarpour
Abstract: The main goal of this study was to investigate the role of stimulus contents on physiological responses. Usually some pauses between the incentives are considered in an experiment to bring the physiological signals to the base. However, the role of these halts on emotional responses has not been carefully evaluated so far. Therefore, in the current study the effect of pauses between emotional blocks of an inducement was inspected. To this effect, autonomic signals, including Electrocardiogram (ECG), Pulse Rate (PR), and Galvanic Skin Response (GSR) were recorded using two emotion elicitation paradigms: image induction and music excerpts. The cease periods between affective blocks of incentives were identified on the signals and implemented on the rest of the procedure. Applying standard and nonlinear based approaches, the following features were extracted: mean, standard deviation, minimum, maximum, median, mode, the second, third, and fourth moment, scaling exponent, fractal dimension, Lyapunov exponent, approximate entropy. The Mann-Whitney U-test was employed to determine the significant differences between the pause segments and the rest. ECG indices for the pause segments are strictly similar to the ones obtained in the baseline measurements. While there is a significant difference between these periods for GSR and PR. These findings are comparable for both music and image paradigms. The results of this study indicate that pause duration between affective blocks of stimuli has a great impact on the emotional autonomic responses. The importance of specialized protocol designing for a specific biomedical signal is concluded from the findings.
Keywords: Signal; Emotion; Image; Music; Pause Intervals.
Towards understanding the etiology of high myopic strabismus using mechanical analysis and finite element modeling
by Haipeng Liu, Yinglan Gong, Zhiqing Chen, Shahina Pardhan, Rajshree Mootanah, Ling Xia, Dingchang Zheng
Abstract: It has been widely accepted that the pathology of high myopic esotropia, a special form of strabismus, is still not well understood. In this study, the mechanical analysis and finite element analysis (FEA) of the oculomotor system were developed from clinical MRI data and applied to examine the physiological hypotheses of extraocular muscle obliquity and deformation respectively. Our mechanical analysis indicated that the muscular obliquity is not the main cause of high myopic strabismus. Next, by simulating the effect of different forces applied to the cross section of each extraocular rectus muscles, the corresponding eyeball rotations were quantified on normal eyes, and high myopic eyes with and without strabismus. The model suggests that the limitation of rotation in high myopic strabismic eyes is mainly caused by the extraocular muscle deformation instead of its obliquity, providing a better understanding of the etiology of high myopic strabismus. To the best of our knowledge, this is the first mechanical and FEA model developed from clinical data to investigate the etiology of high myopic strabismus, providing important tools for future pathological study.
Keywords: eye motion; finite element analysis; high myopia strabismus; mechanical analysis; simulation.
SMOTE and ABC Optimized RBF Network for Coping with Imbalanced Class in EEG Signal Classification
by Satchidananda Dehuri, Sandeep Sathpathy, Alok Jagadev
Abstract: This paper proposes a novel approach for coping with imbalanced class problem by combining the best attribute of synthetic minority over-sampling technique (SMOTE) and artificial bee colony optimized radial basis function neural networks to identify epileptic seizure from electroencephalography (EEG) signal. In brain certain electrical signals have been generated that can be recorded and analyzed for detection of several brain disorder diseases. EEG is the recording of such signals. Careful analysis of the EEG recordings can provide valuable information and understanding the mechanisms of several brain disorder diseases e.g., epileptic disorders. The main feature of epilepsy is the recurrent seizures. Since epileptic seizures occur irregularly and unpredictably, automatic seizure detection in EEG recordings is highly required. We have used Discrete Wavelet Transform (DWT) technique for extraction of potential features from the signal. For classification of these signals into two classes, we have trained the RBFN by a modified version of ABC algorithm (MABC). In the modified ABC, the onlooker bees are selected based on binary tournament unlike roulette wheel selection of ABC. Additionally, kernels like Gaussian, Multi-quadric, and Inverse-multi-quadric are used for measuring the effectiveness of the method in numerous mixtures of healthy segments, seizure-free segments, and seizure segments. In this work we realize, this two class classification problem is highly imbalanced i.e., the instances in one class known as majority class outnumber the instances of other class called the minority class. However, this problem is not uncommon in the field of medical diagnosis. The SMOTE is first applied to generate synthetic instances in the positive class to balance the training data set. Using the resulting balanced dataset, the MABC optimized RBF network is then constructed to identify the epileptic seizure. In the case of imbalanced dataset, our experimental outcomes confirm that RBFN with inverse-multi-quadric kernel trained with MABC is significantly better than RBFNs with other kernels trained by ABC and MABC.
Keywords: Electroencephalography; Radial basis function neural networks; Artificial Bee Colony; Discrete wavelet Transform; Imbalanced Class.
FORCE EVALUATION AND STRESS DISTRIBUTION AT POSSIBLE WEIGHT AND STRUCTURE OF FEMUR BONE IN PELVIS FRAME WHILE STANDING
by Bhaskar Kumar Madeti, Chalamalasetti Srinivasa Rao, S. K. Sundara Siva Rao Bollapragada
Abstract: Abstract: The Present paper gives a clear idea of hip joint, it is very complicated structure here we resolve the forces acting on hip joint by using lami‟s theorem, by using the geometry from CT Scan the tensile force of ligament and axial compressive forces across the femoral head are obtained at various possible positions of pelvis frame while standing. Analysis is done for a person whose weight varies from 600N to 1500N while standing, it is also extended to finite element analysis, stress distribution and deformation is noticed on the surface of femur. Factor of safety is also obtained.
Keywords: keywords: Hip joint; Femur ,Acetabular cup; Axial compressive force; Tensile force; Ligament; Stress distribution; Deformation.
ACTIVITY MONITORING AND MEAL TRACKING FOR CARDIAC REHABILITATION PATIENTS
by Usama Pervaiz, Saed Khawaldeh, Tajwar Abrar Aleef, Vu Hoang Minh, Yeman Brhane Hagos
Abstract: Heart disease has been ranked as the first cause for death in Pakistan along with being the leading cause of death worldwide. Recovering heart patients need to monitor their diet and activity routines to mitigate chances of another occurrence. Our application aims to provide a platform that would help patients to keep track of their activity, alongside allowing the patients to keep track of their diet; which will act as fitness monitoring for cardiac rehabilitation patients. Further more, Our activity classification algorithm by using only seven features shows classification into five different classes with an accuracy of 98\% in real time. Also, Our Meal tracking module is designed to support the cardiac patients with timely feedback and notifications about their dietary routine.
Keywords: Cardiac Rehabilitation;Heart Diseases;Meal Tracking;Activity Monitoring; Support Vector Machine;Scikit Toolbar;Machine Learning;Android application;real time monitoring.
SEPHYRES 2: Applying semantic-pseudo-fuzzy methods in medical diagnostic ontologies
by Ali Sanaeifar, Mahmood Tara, Ahmad Faraahi, Bibimasoumeh Mir Mousavi, Mitra Ahadi, Ali Bahari
Abstract: To date, ontology-based medical diagnostic systems have not incorporated complete descriptions of diseases and their semantic relations with signs and symptoms. In SEPHYRES 1, a pain-focused-only solution was proposed which applied not only general semantic reasoners, but also weight spreading techniques. Proceeding the research, we developed the SEPHYRES knowledge base to address all signs, symptoms and complex relations, including similar terms, terms with the variant generality level, composed terms, terms which include several other terms based on medical diagnostic criteria. The evaluation outcomes, in terms used in patients descriptive history in both of the MEDSCAPE and PubMed case studies, showed that the recall amount of system-oriented evaluation was about 90% provided that just top ten results were considered. Furthermore, the Wilcoxon signed-ranked test between SEPHYRES 2 and the best symptom checker, Isabel engine power, showed that the SEPHYRES 2 significantly improved the matching process of the patient's disease profiles.
Keywords: Computer assisted diagnosis; Clinical decision support; Knowledge modeling; Computer assisted decision making.
Parametric Electrical Impedance Tomography for Monitoring Bone Mineral Density in the Spine Using 3D Human Model
by Neta Naimark, Shimon Abboud, Marina Arad
Abstract: Monitoring methods of bone mineral density (BMD), the standard measure for osteoporosis diagnosis, are both costly and complex. Since changes in bone permittivity and conductivity values occur due to changes in BMD, they can be used as a simple and inexpensive tool for monitoring BMD. In this work the parametric Electrical Impedance Tomography (pEIT) method for monitoring BMD in the spine using 3D human model is theoretically evaluated. Numerical solver on the forward problem in 3D is used for computing electric potential measured on body surface. Varied spinal BMD are simulated by varying bone relative permittivity and conductivity values which represent different disease stages. The inverse problem is solved by creating a lookup-table of different BMD values.
Keywords: Spine bone mineral density; Electrical impedance tomography; Osteoporosis; Simulation Study.
Breast abnormality based early diagnosis of breast cancer using noninvasive digital infrared Thermal imaging
by Priya Hankare
Abstract: Breast Thermography is considered particularly valuable for early breast tumors detection. The fast growing tumor has a higher metabolic rate and associated increase in local vascularization. It will cause the occurrence of some asymmetric heat patterns. Clinical interpretation of a breast thermogram is primarily based on the asymmetry analysis of these heat patterns visually and subjectively. In this paper, a new approach early detection of breast cancer is proposed using asymmetry analysis of breast Thermograms. The heat patterns are first segmented with mathematical morphology. The asymmetry analysis is performed by using histogram generation and feature extraction. The abnormality of a breast thermogram is clearly indicated by the features extracted.
Keywords: Breast Cancer; Infrared Thermal Imaging; Thermography; Asymmetry Analysis; Feature Extraction.
Hard Exudate Based Severity Assessment of Diabetic Macular Edema from Retinal Fundus Images
by Deepthi K Prasad, L. Vibha, K.R. Venugopal
Abstract: Diabetic macular edema (DME) is a consequence of diabetic retinopathy characterized by the abnormal accumulation of fluid and protein deposit in the macula region of the retina. Prior disclosure of even a trivial trace of DME is essential as it could consequently lead to blurred vision. DME can be diagnosed by the presence of exudates (glossy lesions) in the retinal fundus images. In this work, OD and macula are detected using morphological operation and hard exudates are segmented. Exudates are classified using Early Treatment Diabetic Retinopathy Standard as normal, moderate or severe cases. The proposed work also incorporates the extraction of various features from the retinal fundus image. Various textural and exudate features are extracted and fed to a classifier to detect DME. Experiments are performed on a publically available database. Performance is evaluated with metrics like accuracy, sensitivity, specificity and accuracy, the results obtained are promising.
Keywords: Diabetic macular edema; macula; optic disc; hard exudates; feature extraction; classification; Random forest.
In-depth Analysis of Neural Network Ensembles for Early Detection Method of Diabetes Disease
by Bayu Adhi Tama
Abstract: Lifestyle-driven disease such as diabetes mellitus has become a serious health problem worldwide. We propose the fusion of neural network-based classifiers, i.e. neural network and support vector machine to assist in early detection of diabetes mellitus. These classifiers are combined to produce the final prediction. However, when considering a number of classifiers in the pool, the selection of combination rule is not easy to understand. The aim of this paper is to investigate the performance of different combination rules, including several single classifiers involved in the ensemble. We use various performance metrics and validation tests to assess the performance of these classifiers using a real-world dataset. Finally, among the classifiers we evaluate their performance differences using statistical significant test. The experimental results indicate that combination rule with average voting scheme is the best performer compared with other combination rules and single classifiers in the ensemble.
Keywords: Diabetes mellitus; Neural-based classifiers; Classifier Ensembles; Early Detection.
Neuro-fuzzy Implementation for Cervical Lesions Screening in Commercial Sex Workers
by Efosa ODIGIE, Peter Achukwu, Moses Bello
Abstract: The present study proposed a neuro-fuzzy model for cervical lesions (CL) diagnosis. The model intends to use part of reliable preclinical information as a tool to discriminate between abnormal and a normal CL with intent to help the clinicians to have a clear cut of the diagnosis. Therefore, our aims are to formulate a model that is understandable, practicable and capable of predicting CL accurately and use the model in the screening of commercial sex workers (CSWs) operating in some rural communities of Edo state. The specific objective is to confirm the levels of precision of the neuro-fuzzy model the using liquid-based cytology (LBC) method. Adaptive Neuro-fuzzy Inference System (ANFIS) model formulation and liquid-based cytology (LBC) techniques were used for confirmation of ANFIS reliabilities and outcomes. The classification performance of ANFIS model had 98.7% precision with a training error of 1.1652 at epoch 20 on the training dataset and basic testing error 1.255 and 100% accuracy. ANFIS confirmation by LBC showed that 8 cases of CL (15.4%) exist within the confined of 14 brothels (1 case of a low-grade squamous cell intraepithelial lesion, 2 cases of acute and 5 cases of chronic cervicitis), and was age and duration of commercial sex-dependent (P<0.05). ANFIS implementation in this study is excellent for routine screening of the CSWs, while LBC remains a gold standard for CL diagnosis. ANFIS model may be a useful approximation tool in making faster decisions for cervical lesion screening in commercial sex workers.
Keywords: ANFIS; Cervical Lesions; Commercial Sex Workers; Liquid-Based Cytology.
Classification of heart rhythm disorders using instructive features and artificial neural networks
by Santanu Sahoo, Priti Das, Prativa Biswal, S.K. Sabut
Abstract: Accurate detectionof the heart rhythm disorders at an early stage is helpful for enhancing survivable rate. This paper presents an automated detection and classification process of cardiac arrhythmia by time-frequency analysis of the recorded ECG signals from the MIT-BIH arrhythmia database.The discrete wavelet transform (DWT) has been used effectively to remove noise components in order to enhance the ECG signal. The Hilbert transform with the adaptive thresholding technique has been applied to find precise R-peaks and extracting discriminating features in the ECG signal classification. The temporal, morphological and statistical features have been extracted from each heartbeat and then used as input to train the neural network classifier. The NN classifier such as multilayer perceptron (MLP-BP), RadialBasis Function (RBF-NN) and probabilistic neural network (PNN)has been used to discriminate five important types of cardiac arrhythmia beats. The evaluated result of the proposed method shows the best detection performance with sensitivity of 99.92% and positive predictivity of 99.91% with least detection error rate of 0.17% for detecting QRS complex. The classification process produced results with an average accuracy of 98.72%, 99.77% and 99.16% in MLP-BP, RBF-NN, and PNN respectively. The result indicates that the efficiency and the robustness of the proposed methodto detect the precise R-peaks, QRS complex and classifying ECG beats that provides useful physiological information for diagnostic of heart rhythm disorders.
Keywords: Heart rhythm,discrete wavelet transform; Hilbert transform; adaptive thresholding; QRS complex; NN classifier.
Analysis of Texture for Classification of Breast Cancer on Mammogram Images
by Hanung Adi Nugroho, Hanifah Rahmi Fajrin, Indah Soesanti, Ratna Lestari Budiani
Abstract: Breast cancer is a top cancer among women in the world. In conventional methods, medical experts based on observation on patients mammogram images can detect breast cancer. In some cases, this method could lead to misdiagnose in distinguishing an interest object with naked eyes due to the low quality of images. This research aims to classify mammogram images into three classes, i.e. normal, benign and malignant classes based on texture features. The input image is a full part of image without cropping or segmenting the region of interest (RoI). Some pre-processing techniques were involved, including eliminating label, removing the artefacts, cropping breast area, contrast enhancement and smoothing with median filter. Afterwards, some texture features were extracted followed by classification process by using multi-layer perceptron (MLP) classifier with back error propagation (BEP) learning. The performance of proposed method is evaluated using 60 mammogram images taken from Mias dataset. Classification result gained a maximum value with contrast limited adaptive histogram equalisation (CLAHE) for contrast enhancement method, GLCM-based for feature extraction method and MLP as the classifier with the accuracy of 98.33%, sensitivity of 100% and specificity of 97.50% for both normal and abnormal classes. On the other hand, the classification of three classes (normal, benign and malignant) obtained an accuracy of 90%, sensitivity of 85% and specificity of 87.50% directly without any hierarchy. These results indicate that the proposed method successfully classifies the abnormality of mammogram images.
Keywords: breast cancer; texture feature; CLAHE; grey level co-occurrence matrices; multi-layer perceptron.
A Distributed Integration System Enabling Electronic Health Records: An Italian Experience
by Cinzia Muriana, Giulio Gilia, Vincenzo Mistretta, Tommaso Piazza, Giovan Battista Vizzini
Abstract: Today distributed systems are commonly employed in patient data management to improve the care process. Moreover, distributed systems can be usefully applied to realize lean and flexible architectures that facilitate health care structure interoperability. Systems used in health care structures are characterized by proprietary data flow formats and encoding systems, which hinder the possibility of sharing data in a standard format, and extracting atomic data for further analysis. This paper describes the architecture of a distributed interoperability system able to structure flow data into standard formats and send them to an advanced Electronic Health Record designed as part of the Smart Health 2.0 project (PON04a2_C). The use of the distributed interoperability system within the experimental phase of the Electronic Health Record highlights the flexibility of the proposed solution in meeting the reality of the health care structures involved.
Keywords: distributed systems; informatics health care solutions; CDA2; electronic health records.
A Novel Patient Friendly IT Enabled Framework for Selection of Desired Health Care Provider
by Salaja Silas, Elijah Blessing Rajsingh
Abstract: The healthcare providers have tremendously grown during the past decade. Selecting a healthcare provider based on patients preference is really a great challenge. The purpose of this paper is to provide an IT enabled framework that will help the patients to find a suitable healthcare provider based on their preferences. The patients preferences in selecting a healthcare provider differ based on their geographical region, gender, age and socio-economic background. Discovering a set of feasible healthcare providers and selecting the most appropriate healthcare provider based on the patient preferences can be modeled as a multi-criteria decision making problem. In this article, a novel patient friendly framework is proposed for selecting the best healthcare provider. Even with a huge data set, the results show that the proposed framework is robust, reliable, faster and effective. The analysis reveals that the number of preferences and the healthcare providers influence the healthcare selection time. Case studies on 500 patients from different geographical regions reveal that the proposed framework is effective.
Keywords: Healthcare provider; multi-criteria decision making; service selection; ELECTRE.
Real-Time Signal Processing of Photoplethysmographic (PPG) Signals to Estimate the On-Demand and Continuous Heart Rate by spectral Analysis
by Madhan Mohan, Nagarajan V, Vignesh J.C
Abstract: In Healthcare applications, heart rate is one of the vital signs which give the health informatics of a person and also it is an important parameter to monitor for athletes and fitness enthusiasts during their workouts. In earlier days, the Heart Rate monitoring devices like ECG are very expensive and required more hardware resources. But nowadays, the evolution of PPG sensors help to develop low cost Heart Rate monitoring devices with minimal hardware resources. PPG sensor based Heart rate monitoring devices are non-invasive low cost devices and it can be used in real time applications to estimate the heart rate of a subject. In this paper, an efficient algorithm is proposed to find the Heart rate using frequency spectrum analysis on PPG signals. The proposed algorithm uses minimum hardware so it can be targeted in real time embedded platforms. Using the proposed algorithm, the Heart rate is calculated with a pass percentage of 99.2, Mean Absolute Percentage Error (MAPE) of 1.59%, Mean Absolute Error (MAE) of 1.20 BPM and Reference Closeness Factor(RCF) of 0.989. The First reliable Heart rate output from the algorithm comes in 6.5th second, which is the minimum possible time. The algorithm operates with a speed of 2 MIPS and with a memory of 18 KB. So the proposed method can be integrated to any low cost real-time embedded platforms to accurately measure the Heart Rate.
Keywords: Heart rate; Photoplethysmography (PPG); Frequency domain; Smoothing; Decimation; FFT.
Detection of Abnormal Blood Cells by Segmentation and Classification
by Abdellatif Bouzid-Daho, Mohamed Boughazi, Eric Petit
Abstract: Leukemia is a cancer of the hematopoietic cells. The detection of abnormal cells before cancer degeneration is a medical problem. The aim of our work is to obtain maximum recognition rate of leukemia. We propose the development of a system based on mathematical morphology and k-means methods capable of segmentation, classification, and detection of the cancerous blood cells. This allows the characterization and the description the cancerous region, which is an important task in the interpretation and diagnosis of pathologies present in blood. The segmentation was carried out using an efficient and fast algorithmic processing. It turns out that the proposed system shows to better segmentation and classification for tested images. The obtained experimental results are very encouraging which help hematologists for identification of abnormal blood cells.
Keywords: Leukemia; k-means; segmentation; classification; diagnostic; abnormal blood cell.
Study of Speech Enabled Healthcare Technology
by Saswati Debnath, Pinki Roy
Abstract: Cost and quality of healthcare are the most challenging requirements in todays fastest growing medical technology and to meet these requirements Automatic Speech Recognition (ASR) is one of the blessings to the medical world. Automatic speech recognition offers the potential to dramatically improve the cost and quality of healthcare service; many developments took place in national and international standard like e-prescription, clinical documentation, speech recognition in radiology, pathological speech signal analysis etc. Speech based healthcare application also helps to access remote healthcare service. Recently large number of hospitals and doctors use these types of medical technology to deliver better services to the patient. This paper presents a review of recent advancement on medical technology based on speech recognition. It covers some of the speech based healthcare research and software in national and international standard, the implementation framework and application and also this paper come up with a new idea that is identify the symptoms of diseases using machine learning. The symptoms are given through speech. The experimental result shows the accuracy of identifying the symptoms through speech recognition. The aim of the research is to improve medical technology using machine learning and speech recognition technology. In future it will help people for remote health care where doctor will not physically available and reduce workload pressure of doctor.
Keywords: Medical technology; Automatic speech recognition (ASR); Speech features; Machine learning.
Automatic Ostia Detection in CTA Volume Data: A Comparative Study
by Noha Seada, Mostafa G.M. Mostafa
Abstract: Objectives: Automatic coronaries ostia detection is implemented using two different methodologies: Template Matching and Corner Detection. Methods/Analysis: The proposed methods used for the ostia detection are selected for implementation after studying ostia anatomical features and observing their appearance on a segmented ascending aorta contour. Findings: Template matching is implemented for coronaries ostia detection, since they have a special pattern on a segmented ascending aorta contour. Therefore two templates are identified, one for the right and one for the left coronaries ostia. These two templates are matched across the input volume to detect the coronaries ostia. For corner detection; Harris corner is implemented to detect coronaries ostia as corners appearing on the ascending aorta. It is applied on the volume to detect corners and ostia points are correctly located based on ostia anatomical features that automatically identifies the ostia points. Novelty/Improvement: The two methods used for automatic ostia detection succeeded to detect the coronaries ostia points in all test cases and results are validated versus a ground truth. Although both approaches succeeded in ostia detection; the template matching accuracy and processing time are better than that of the Harris corner detection. Moreover, the two methodologies presented for comparison; give competitive accuracy and processing time with respect to other previous work for ostia detection and this proves the correctness and efficiency of the methodologies.
Keywords: Template Matching; Automatic Ostia Detection; Computed Tomography Angiography (CTA).
Learn Quest A Virtual Reality based system for training Autistic Kids
by Shriram K Vasudevan, Abhishek S N, Swathi S, Akshaya Lakshmi, Anandram S
Abstract: Childhood is the best part of life. A child enjoys anything and everything in this world without any discriminations. It is also the period of rapid brain development, where the kids learn the stepping stones for life. Autism Spectrum Disorder (ASD) is a group of developmental disabilities that causes major impairments in social, communication and behavioral aspects of the children. Unlike other children, an autistic kid finds it difficult to go along and communicate with others. Social interactions and communication are the major aspects for peaceful living of any human being, and these children face difficulties in the same. So, we have developed a virtual reality based game engine that would attract the children and make them enjoy interacting with it while increasing the chance for learning. As the children continues playing this game, their learning and interacting ability increases nullifying some effects of autism, thus making it a boon for these kids.
Keywords: Virtual Reality; Autism; 3D; game; interaction; Gesture; Leap Motion Sensor.
Mechanics of the septal wall may be affected by the presence fibrotic infarct in the free wall at end-systole
by Fulufhelo Nemavhola
Abstract: The relationship between infarct mechanics and ventricular function of septal wall in the presence of fibrotic infarct in the free wall remains poorly understood. This study compares the mechanics of the septal wall in the healthy and fibrotic infarcted rat hearts models during active ejection. Finite element models of rat heart were developed to study the ventricular mechanics of the septal wall in the presence of fibrotic infarct in the left ventricular free wall. To simulate active contraction, three dimensional constitutive Fung model was implemented in Abaqus
Keywords: septal wall; myocardial mechanics; heart computational models; fibrotic infarct; infarct modelling.
Optimized DBN for Effective Enhancement of Ultrasound Images with pelvic Lesions
by Sandanand L. Shelgaonkar, Anil B. Nandgaonkar
Abstract: Nowadays, the ultrasound modality is the current research areas for lesion analysis. Moreover, it is renowned; the ultrasound images exhibit hassle- free utilization as well as cost effectiveness. In the literature, merely a small number of reliable contributions have been reported from the females for analyzing the pelvic lesions. However, the literature lags with adopting a prominent enhancement procedure to facilitate the diagnosis process. This paper adopts an Optimized Deep Belief Neural (ODBN) network for enhancing the US image of pelvic portions. It considers the higher order as well as lower order statistical characteristics of the image to define the appropriate filter band for image enhancement. The lower order features are optimized to use with the higher order features so that the optimal usage of features is reported here. To optimize the lower order features, an advanced optimization search algorithm named Grey Wolf Optimizer algorithm (GWO) is exploited. The ODBN learns the optimized features and the noise characteristics for precise prediction of the filter bands, which enhance the image substantially over the conventional filter bands. The performance of the proposed method is compared with the conventional methods using the benchmark as well as real-time US images of pelvic lesions. The quality of enhancement is ensured using the renowned measures such as PSNR and ESSIM that exhibit the performance of the proposed approach.
Keywords: ultrasound; PSNR; ESSIM; GWO; DBN.
An innovative technical solution to avoid Insomnia and Noise-Induced Hearing Loss
by Shriram K V, Ikram Shah, Sriharsha Patallapalli, Karthikeyan S, S. Subhash Chandran, U. Adithya Bharadwaj
Abstract: The major challenge these days with the increased usage of the mobile phone is loss of sleep (insomnia), increased stress and finally more damages to the health and mental wellness. Most of us have the habit of even keeping the phone underneath the pillow while sleeping. There are people who are now treated for mobile addiction. People are there to use a phone to get sleep sooner. Sleep time is meant for the brain to rejuvenate, organize thoughts and relax. If the sleep patterns are disturbed due to a continuous external commotion, it will result in the subconscious to spend energy and brain space. Hence, it is wise to switch off the music after the person sleeps which most of us do not do as we are already slept by then. Also, wearing headphones for an hour will increase the bacteria in your ears by 700 times. Many researches in the past prove that music is a solution to reduce stress and to bring sleep. But, we, the users are using headphones for music and it causes some dangerous side effects including affecting the sleep and noise-induced hearing loss. There is a strong need to avoid unnecessary health problems which are meant to be avoided. Here, we propose a system which will ensure that the music player is stopped after understanding that the person using it has slept and he no more needs the tunes thereby not disturbing the deep sleep and preventing insomnia and noise-induced hearing loss. Novelty is what we present in this research through making sure if the person has really slept or not. Our innovative system switches off the audio automatically when a person falls asleep. The same can be expanded and improvised to provide analytics to the user as in how many hours did he sleep peacefully, how is his body tuned towards listening to music, how long does he need music, under what conditions does he over listen or under listen etc.
Keywords: Sleep monitoring; Headphones/ Earphones ; Wearables; Heart rate ; Pulse; Mobile Application; Multimedia Playing; Insomnia; NIHL.
Health Panel: a Platform Useful to Physicians for Fast and Easy Managing of FPGA-based Medical Devices
by Agostino Giorgio
Abstract: The Design of Medical Devices is frequently based\r\non Field Programmable Gate Arrays (FPGAs). This is especially\r\ndue to their unique properties in order to fast prototyping and\r\ntheir great capabilities in Digital Signal Processing (DSP). The\r\nproblem arises to make physicians able to manage such devices\r\nespecially to set and program during prototyping, experimental\r\nand validation processes of new medical devices. Therefore, the\r\nobjective of this paper is the description of a platform, named\r\nHealth Panel, in its hardware and software components,\r\novercoming these issues. The design methods and the test and\r\ndebug methods are also detailed. The hardware is an Embedded\r\nSystem (ES) and the software, developed in Matlab environment,\r\nacts as user friendly interface making the physician able to\r\nmanage quick and easy the FPGA. Results of tests and of an\r\naccurate debug of the platform are described. The platform aims\r\nat providing a significance contribution for designing hardware\r\nand software interfaces for an easy and quick use of FPGA-based\r\nmedical devices. Conclusions and final remarks provide\r\nsuggestions for future developments of the platform.
Keywords: FPGA; Embedded Systems; Digital Signal
Processing; Matlab; Medical Devices; Electrocardiogram,
Spectro-Temporal Analysis of Electromyogram Signals (EMGs)
by Meziani Fadia, Rerbal Souhila, Debbal Sidi Mohammed El Amine
Abstract: Electromyography (EMG) signals can be used for clinical/biomedical applications. Is a very important biomedical signal that measures electrical currents generated in muscles during its contraction representing neuromuscular. Individual muscle fiber action potentials are sometimes acquired using wire or needle electrodes placed directly in the muscle. The combination of the muscle fiber action potentials from all the muscle fibers of a single motor unit is the motor unit action potential (MUAP) which can be detected by a skin surface electrode (non-invasive) located near this field, or by a needle electrode (invasive) inserted in the muscle. The simple detection with electrodes has been recognized for a long time as an important tool for the diagnosis of neuromuscular diseases, although its accuracy is still insufficient to diagnose some neuromuscular diseases (Myopathy and Neuropathy). It does not enable the analyst to obtain both qualitative and quantitative characteristics of the EMG signals. The efficiency of diagnosis can be improved considerably by using modern digital signal processing techniques. Therefore, these last can provide useful and valuable information on these signals. The aim of this study is to analyse EMGs signals using temporal and frequency analysis. This analysis is based on temporal parameters such as :amplitude, energy, entropy... and frequency parameter such as median frequency and mean frequency by using the FFT and the non-linear information by using the bispectrum analysis , can provide a wide range of informations related to the type of signal (normal and pathological).
Keywords: Electomyogram; EMG; signal; time-frequency; analysis; parameters; pathology.
Generating a stable primary schedule for an integrated surgical suite
by Asie Soudi, Mehdi Heydari
Abstract: Efficient utilization of operating room (OR) is a common anxiety of surgical suit manager which necessitate an effective planning and scheduling of surgeries. In this paper we investigate predictive / reactive scheduling of an integrated surgical suite in the form of a two-stage hybrid flow shop scheduling problem (HFSP). The deterministic model comprises both assignment and sequencing decisions for elective surgeries. By further considering shared capacity between elective and emergency patients, a chance constrained programming model is extended for the first time to cope with uncertain disruption. It is shown that how a chance constrained model will reduce to just considering an augmented surgery which processing time depends on distribution function of emergency surgery processing time and confidence level of scheduler. Two new important measures in reactive scheduling literature, 'stability' and 'robustness' are taking into account in surgical suite scheduling for the first time. Computational results demonstrate the efficiency of primary schedule generated by extended chance constrained programming model as well as the effectiveness of new measures in hospitals. As the chance constrained model is NP-hard, a decomposition heuristic algorithm based on tabu search (TS) is proposed to cope with problems of real size.
Keywords: Stability and robustness; predictive / reactive scheduling; hybrid flow shop; integrated surgical suite.
Adaptive Improved Binary PSO Based Learnable Bayesian Classifier for Dimensionality Reduced Microarray Data
by Barnali Sahu, Satchidananda Dehuri, Alok Jagadev
Abstract: This article presents, an adaptive improved binary particle swarm optimization based learnable Bayesian classifier for dimensionally reduced microarray data. In the first fold of this two-folded work, the problem of dimension is reduced by unsupervised method of feature reduction. We apply k-means clustering algorithm on the microarray data to group functionally redundant genes followed by application of signal-to-noise-ratio ranking technique to generate an intermediate feature subset consisting of most relevant and non-redundant feature subsets. In the second fold, the feature subset has been given to adaptive binary particle swarm optimization based learnable Bayesian classifier for simultaneous selection of features and classification. We conduct an extensive experimental work on a few benchmark datasets to validate its classification accuracy with and without reducing the dimensionality of micro-array data. It was observed that our method is not only accepted as a good classifier over methods, which are considered here for comparison but also be treated as an alternative method of reducing dimension of the problem.
Keywords: Microarray data; Feature selection; Signal-to-Noise-Ratio; Classification; Bayesian classifier; Adaptive PSO.
Product Unit Neural Network trained by an Evolutionary Algorithm for diabetes disease diagnosis
by Radhwane BENALI, Nabil Dib, Fethi Bereksi Reguig
Abstract: Diabetes disease occurs when the level of glucose in the blood becomes higher than normal because the body is unable to produce the insulin which is needed to regulate glucose. In this study, a new classification method for the diagnosis of diabetes disease was developed. This method is based on a special class of neural network known as product-unit neural networks (PUNN) which was trained by an evolutionary algorithm (EA). We have used EA in order to determine the basic topology of the structure of the PUNN, and to estimate its coefficients weights. The performances of the proposed classifier were evaluated through the sensitivity, the specificity and the classification accuracy using both conventional and 10-fold cross-validation method using the Pima Indian Diabetes (PID) dataset. Obtained results reveal that the proposed approach outperforms several famous and recent methods existing in the literature for diabetes disease diagnosis.
Keywords: Product Unit Neural Network (PUNN); Evolutionary Algorithms (EA); Diabetes disease diagnosis; Pima Indian diabetes (PID).
Development of a Serious Game to Enhance Assistive Rehabilitation
by Mauro Bellone, Andrea Martini, Francesco Argese, Piero Cirillo, Italo Spada, Antonio Cerasa
Abstract: The aim of this study is the investigation of novel assistive technologies based on serious gaming for the assessment of postural control and motor rehabilitation. Previous research already demonstrated that rehabilitation, assistive technologies and physical activities can improve patients' quality of life, and virtual reality applications may act as good additional companions during the therapeutic sessions. Indeed, workout routines supported by serious gaming encourage patients to train harder, bringing the therapy towards a pleasant game into a virtual environment, where specific goals must be reached. During the game, sensors track patients' movements and transfer the data to software components that record a database from which patient progress can be determined.
Acquired measures have been interpreted in the form of new biomarkers, which enable the assessment of postural control. Such biomarkers are based on a probabilistic approach and show the capability to discriminate between well-performed exercises and incorrect movements. A long-term experimentation on a specific exercise is proposed, showing a considerable improvement in the patients' performances ranging between 5\% and 30\%.
The main contributions of this research are:
i) the definition of new biomarkers for the postural assessment of patients affected by motor disorders;
ii) the use of non-intrusive technologies, which enhances the freedom of movement for patients, increasing the results reliability;
iii) the design of a virtual reality interface, which allows patients to interact in a pleasant and familiar environment without constant supervision;
iv) the development of a new editor to easily customize virtual exercises, analyze rehabilitation progress, and create statistics.
Keywords: Assistive technologies; rehabilitation; virtual reality; motor inabilities; visual tracking.