International Journal of Image Mining (9 papers in press)
Trigonometry Based Motion Blur Parameter Estimation Algorithm
by Ruchi Gajjar, Tanish Zaveri, Asim Banerjee, K. V. V. Murthy
Abstract: Restoration of blurred images requires information about the blurring function, also known as point spread function (PSF), which is generally unknown in practical applications. Identification of blur parameters is essential for yielding blurring function. This paper proposes a technique for estimation of parameters of motion blur by formulating the trigonometric relationship between the spectral lines of the motion blurred image and the blur parameters. In majority of the existing motion blur parameter estimation approaches, the length is estimated by rotating the spectrum to the estimated angle. This requires the angle estimation to be done forehand. The proposed method estimates both, length and angle simultaneously by deriving the trigonometric relation between spectral lines, thereby eliminating the need of rotating the spectrum for length estimation. The proposed technique is applied on standard database Berkeley segmentation dataset, Pascal VOC 2007 dataset and USC-SIPI Image database. The simulation results show that length and angle are accurately estimated by the proposed method. The performance of proposed method is compared with existing state of art techniques which show that proposed method exhibit better performance in terms of the test range and parameter estimation.
Keywords: image degradation; motion blur; parameter estimation; point spread function.
A Novel Method for Query Based Image Retrieval using Prototype Based Clustering
by R.Tamil Kodi, Roseline Nesa Kumari, Maruthu Perumal
Abstract: Content based image retrieval is a process applied for searching of images from a large database in which searching is done based on the content of the image. The content of the image can be of color, texture, and shape. This paper concentrates the primitive feature of texture in which the extraction of image content is considered. The implementation of such a system requires the extraction and storing of the image features to be compared with the features of the query image with this flow, the implementation process is more dynamic since all features have already been stored somewhere. This paper proposed a new method called Prototype Based Cluster (PBC) where the features with similar values or properties are grouped together to form clusters and comparison is made between these clusters with the database images and the relevant images are retrieved and stored. This method will show good performance when compared with existing ones. The experimental result shows the effectiveness of our proposed method PBC applied on the query image.
Keywords: Clusters; Retrieval; Query image; performance; features; extraction;.
Hardware Implementation of Stereo Vision Algorithms for Depth Estimation
by Nitish Wadne
Abstract: Depth estimation has applications like robot navigation, advance driver assistance systems, 3D movies etc. Depth is represented in terms of disparity map which can be generated using various stereo correspondence algorithms. This paper presents an implementation of semi global block matching algorithm on raspberry pi to estimate the depth from the camera. The algorithm computes the disparity using block wise matching and smoothness constraint. The proposed algorithm is compared with SAD algorithm on personal computer as well as on raspberry pi. The algorithm is also further, evaluated on the standard dataset. The project aim is to detect people in an image and estimate their depth from the camera. The real time implementation of the proposed algorithm uses block size of 21 x21 for images which has resolution of 1280 x 720 P. The algorithm estimates depth with an accuracy of 95%. The system also provides faster processing time to the proposed algorithm.
Keywords: Computer vision; depth estimation; Raspberry Pi; Semi-global block matching.
Special Issue on: ICIA-16 Image Processing and Analysis
Analysis of Diverse Optimisation Algorithms in Breast Cancer Detection
by Senthil Kumar K, Venkata Lakshmi K, Karthikeyan K, JasiyaJabeen A
Abstract: Breast cancer is a widespread problem faced by the women in recent years. It is highly essential to detect the breast cancer at an early stage to save lives. The ultrasound image helps more in the diagnosis and analysis of breast cancer than the mammogram. Image segmentation technique is used to segment the mistrustful masses from an ultrasound image of the breast. Optimization algorithms play a vital role in image segmentation techniques which increases the efficiency and accuracy of image segmentation results. This work focuses on implementation and analysis of various optimization algorithms in detecting mistrustful masses in the given ultrasound image of the breast. In preprocessing the speckle noise is reduced by using the median filter and Gaussian filter and proved that the median filter has better performance than the Gaussian filter in the isolation of speckle noise. The visual appearance of the input image is improved by using adaptive histogram equalization. The preprocessed input image is subject to image segmentation techniques with various optimization algorithms viz. particle swarm optimization, chaotic particle swarm optimization, k-Medoid clustering, fuzzy c Means and k-Means clustering with manual selection of cluster centers. A comparative analysis has been done on the above said algorithms using MATLAB and from the results, it is proved that the chaotic particle swarm optimization algorithm has best result among the others. The accuracy and dice similarity coefficient of the chaotic particle swarm optimisation based method is 93.5793 and 0.8735 respectively. This proves that the chaotic particle swarm optimization algorithm is highly suitable in segmenting the breast ultrasound image.
Keywords: Ultrasound image; Median filter; Gaussian filter; Histogram Equalisation; Particle swarm optimization; Chaotic particle swarm optimization; k-Medoids; Fuzzy c-Means; k-Means clustering and Dice coefficient.
Generating Efficient Classifiers Using Facial Components for Age Classification
by Sreejit Panicker, Smita Selot, Manisha Sharma
Abstract: Aging a natural phenomenon, happens with time and becomes evident as person grows. An individual undergo various changes as age progresses. This is noticeable by his or her facial structure and texture which changes as growth accelerate. Facial growing is a standard happening that is sure, and differs from individual to individual subject on the conditions and living susceptibility. Uses of age assertion are seen in areas like Forensic science, security, and furthermore to decide wellbeing of an individual. Facial parameters used for age characterization can be either structural or textural. In this paper we have used statistical methodologies for feature extraction. In structural, facial development is considered for characterization, by figuring the Euclidean separation between the different points of interest on facial image. The experimental results are significant and remarkable.
Keywords: Aging; Age Estimation; Facial Images.
Object Boundary Detection through Robust Active Contour Based Method with Global Information
by Ramgopal Kashyap
Abstract: Restorative applications have turned into a help to the healthcare industry various therapeutic applications require legitimate segmentation of medical images for an exact determination. These applications guarantee astounding segmentation of therapeutic images utilizing the level set method is a popular method, yet quick handling with exact portions remains a test. The result of the poor radio recurrence loop consistency and inclination driven swirl streams, there is much clamor and intensity inhomogeneity in CT images, and it seriously influences the segmentation exactness, better segmentation results are hard to accomplish with conventional methods. In the proposed method initially composed energy utilization, enhanced cross section Boltzmann method which replaces the partial differential equation settling approach that sets aside such a great amount of time for preparing. An enhanced energy based active contour method with a level set definition which coordinates with local and global energy terms, local term compels to pull the form and limit it to object boundary, determines significant advantages not restricted to, quick preparing, mechanization, invariance of precise CT image portions,. Thus, the global energy fitting term drives the development of form at a separation of the object boundary. The global energy term depends on the global segmentation calculation, which can better catch force data of images than hybrid region based active contour method. Both local and global terms, are commonly acclimatized to build a level set method to portion images with accuracy. The oddity inside our technique is to quickly understand fractional differential conditions in the level set strategy with Boltzmann strategy which utilizes neighborhood mean a quality which empowers it to identify limits. The proposed strategy infers profitable points of interest not stuck simply utilizing speedy process, computerization and right restorative picture portions. The proposed method performs better both subjectively and quantitatively contrasted with other energy based method.
Keywords: Active Contour Models; Hybrid Region Based Method; Intensity Inhomogeneity; Local Binary Fitting; Local Image Fitting; Mumford Shah Model; Signed Distance Function; Variational Level Set Model.
Computer Aided Mammography Techniques for Detection and Classification of Microcalcifications in Digital Mammograms
by S.Punitha Stephen, Ravi Subban, Anousouyadevi M, Vaishnavi J
Abstract: Recent techniques that are developed in Computer-Aided Mammography (CAM) produce more accurate results in detection and diagnosis of microcalcifications in its earlier state that can lead to breast cancers among women. These techniques aim at the reduction of false positive rates through which the number of biopsies and surgeries can be greatly reduced. This paper gives a detailed study of the existing techniques available in computer-aided mammography for the segmentation and classification of the microcalcifications present in the Digital mammograms which help the radiologists to take quick and accurate diagnosis decisions.
Keywords: Segmentation; Classification; Computer Aided Mammography (CAM); Microcalcification Clusters (MCC); Receiver Operating Characteristics (ROC); Artificial Neural Networks(ANN).
Special Issue on: Medical Imaging
DECISION BASED FUZZY LOGIC APPROACH FOR MULTIMODAL MEDICAL IMAGE FUSION IN NSCT DOMAIN
by Sivasangumani Selvaraj
Abstract: Image fusion is to reduce uncertainty and minimize redundancy in the output while maximizing relevant information from two or more images of a scene into a single composite image that is more informative and is more suitable for visual perception or processing tasks like medical imaging, remote sensing, concealed weapon detection, weather forecasting, biometrics etc. Image fusion combines registered images to produce a high quality fused image with spatial and spectral information. The fused image with more information will improve the performance of image analysis algorithms used in medical applications. In this paper we proposed an image fusion algorithm based on decision approach and NSCT to improve the future resolution of the images. In this, images will be segmented into regions and decomposed into sub-images and then processed using Fuzzy Logic, the information fusion is performed using these images under the certain criteria such as Non Subsampled Contourlet Transform (NSCT) and certain fusion rules such as Fuzzy Logic, and finally these sub-images are reconstructed into the resultant image with plentiful information. The various metrices Entropy, Mutual Information (MI) and Fusion Quality are calculated to compare the results. The proposed method is compared both subjectively as well as objectively with the other image fusion methods. The experimental results show that the proposed method is better than other fusion methods and increases the quality and PSNR of fused image.
Keywords: Image Fusion; DWT; NSCT; Fuzzy logic; Genetics algorithm; Entropy; MI.
Comparative Study of Different Machine Learning Classifiers for Mammograms and Brain MRI Images
by Poonam Sonar, Udhav Bhosle, Chandrajit Choudhury
Abstract: Abstract: Today, Breast cancer in women has become a leading cause of cancer deaths. Till date, Mammography is the most reliable and accurate technique for early and accurate detection of breast cancer. Therefore, the researchers are giving highest priority to computer aided diagnosis of breast cancer through mammograms. This paper presents the machine learning based mammogram classification techniques. Mammogram database images are pre-processed to extract region of interest. GLCM (grey level covariance matrix) based texture features are extracted from segmented ROI. These features are used to trained classifier. The trained classifier is used to classify breast tissues in normal / abnormal class and further to benign / malignant class. Different machine learning classifiers such as SVM, KNN, Random Forest 4.5, Logistic Regression, Fisher Discriminant analysis, Na
Keywords: Keywords: Mammograms; Texture feature ; SVM; KNN; Hybrid SVM_KNN; Random Forest.