Shrinivas D. Desai

Detection of Microcalcification in Digital Mammograms by Improved-MMGW Segmentation Algorithm

Breast cancer represents the most frequently diagnosed cancer in women. In order to reduce mortality, early detection of breast cancer is important, because diagnosis is more likely to be successful in the early stages of the disease. This paper presents an improved multi-scale morphological gradient watershed segmentation method for automatic detection of clustered micro calcification in digitized mammograms. We use adaptive median filter for preprocessing and incorporated corrections after watershed segmentation by cloned data. This correction has led to better detection and localization of micro calcifications. By comparing our results with original multiscale morphological watershed segmentation method, we proved that the proposed technique is better and performance is improved by approximately 20%. The true positive rate and false positive rate are used to evaluate the performance of the proposed technique. For experimental purpose the dataset from Mammographic Image Analysis Society database and few data collected from local diagnostic center. The result shows achievement of true positive rate of about 95.3% at the rate of 0.14 false positive per image.

Back-Propagation Neural Network Versus Logistic Regression in Heart Disease Classification

Globally, cardiovascular (heart) diseases are the major cause of death. About 80% of deaths are reported in developing countries. Looking at the trend and lifestyle, one can predict that by 2030 around 23.6 million people may die due to heart disease (mainly from heart attacks and strokes). Each and every healthcare unit generates enormous heart disease data which unfortunately are not “mined” to discover pattern and knowledge for effective decision making. Practical knowledge by domain experts plays vital role. However, there is a need for effective analysis tools to discover unknown relationships and trends in data. Objective of this paper is to assess the accuracy of classification model for the prediction of heart disease for Cleveland dataset. A comparative study of parametric and nonparametric approach in classifying heart disease is presented. Two classification models, back-propagation neural network (BPNN) and logistic regression (LR), are used for the study. The developed classification model will assist domain experts to take effective diagnostic decision. 10-fold cross validation method is used to measure the unbiased estimate of these classification models.

Content-Aware Video Retargeting by Seam Carving

Due to the rapid growth of digital gadgets with various screen sizes, resolutions and hardware processing capabilities, robust video retargeting is of increasing relevance. An efficient retargeting algorithm should not only retain semantic content, but also maintain spatiotemporal resolution of video data. In this paper, the effective seam carving technique for content-aware video retargeting is discussed. Retargeting video is of immense importance as it is frequently played on several gadgets such as television, mobile, tablet, and notebook. The proposed method considers each video frame as an independent image entity and tries to resize it. Our main contribution is a formulation of seam carving using graph cut method. Convention cut techniques fail to defend a meaningful seam. Single monotonic well connected by pixel to pixel is most desirable property in seam carving process. The traditional seam carving method is designed to work based on the minimum energy concept, while ignoring the energy that has been introduced by the operator. To address this issue, we propose a new design criterion in which least amount of energy is introduced in retargeted video.

Spatial Resolution Assessment in Low Dose Imaging

Computed tomography has been reported as most beneficial modality to mankind for effective diagnosis, planning, treatment and follows up of clinical cases. However, there is a potential risk of cancer among the recipients, who undergoes repeated computed tomography screening. This is mainly because the immunity of any living tissue can repair naturally the damage caused due to radiation only up-to a certain level. Beyond which the effort made by immunity in the natural repair can lead to cancerous cells. So, most computed tomography developers have enabled computed tomography modality with the feature of radiation dose management, working on the principle of as low as reasonably achievable. This article addresses the issue of low dose imaging and focuses on the enhancement of spatial resolution of images acquired from low dose, to improve the quality of image for acceptability; and proposes a system model and mathematical formulation of Highly Constrained-Back Projection.

Seeded Watershed Segmentation Based Proteomics for 2D-Gel Electrophoresis Images

Proteome analysis is most frequently accomplished by a blend of two-dimensional gel electrophoresis (2DGE) to separate and visualize proteins and mass spectrometry (MS) for protein identification. Even though this technique is influential, mature, and responsive, questions remain regarding its ability to characterize all of the elements of a proteome. The process of screening for proteins is laborious and protein pattern differences between gel images can be very subtle and tedious to detect by naked eye. Hence there is tremendous need for automatic detection of proteins by computer based tools. In this paper we propose software tool, based on watershed segmentation and point matching as a promising method for protein detection. Proteomics has become an important part of life Sciences especially after the completion of sequencing the human genome. For the analysis, the software Protein Image Registration (PIR) is used. The proposed tool detects and presents proteins along with their properties such as name, molecular mass and pH score.

User intervention based segmentation of Myocardium In cardiac cine MRI images

For the past one and a half decade, Magnetic Resonance Imaging (MRI) has become a preferred imaging technique for examination of cardiac morphology and its functions in humans. The heart cavities segmentation in MRI is still a challenge to be resolved due to the greater unpredictability of the images among patients and the characteristics of cardiac MR images. We present user intervention based semi-automatic method for segmentation in short axis images. The method includes various segmentation methods like graph-cut, watershed and the threshold based segmentation to calculate wall thickness and ejection factor which are of clinical importance. Challenge is to effectively segment epicardium and endocardium boundaries, for effective assessment. We have collected dataset from Sunnybrook Cardiac Data (SCD). The performance of each of the segmentation method is assessed by recording confusion matrix, and calculating sensitivity, specificity and accuracy. We conclude with a discussion of the results obtained which are in favour of graph-cut segmentation and future intended course of action in this field regarding methodological and medical issues.

Modified HYPR algorithm - A promising technique for low dose imaging

Medical imaging has grown tremendously over years. The CT and MRI are well thought-out to be most extensively used imaging modalities. MRI is less dangerous, but one cannot underrate the unsafe side effects of CT. Current study reveals the actuality of escalating risk of cancer as side effect for patients who go through recurring CT scanning. Consequently the devise of low dose imaging protocol is of the enormous significance in the current scenario. In this paper we present modified highly constrained back projection (M-HYPR) as a most promising method to address low dose imaging. HYPR is basically an iterative process in nature and hence computational greedy, and is the root cause for being neglected by CT developers. The weight matrix module, being main reason for huge computation time is modified in this work. Considerable speed up factor is recorded, as compared original HYPR (O-HYPR) on a lone thread CPU implementation. The superiority of reconstructed image in each platform has been analyzed. The evidenced results convey substantial improved performance by M-HYPR algorithm, and appreciable usage of GPU in medical image applications.

Attainment of Professional Outcome by Active Learning Method

The program outcomes defined in outcome-based education shall be broadly categorized into technical outcome and professional outcome. In engineering education, the professional outcomes are addressed to some extent by introducing humanity science course. Professional outcomes are addressed by mini and capstone projects. However in core courses the activities need to be designed to address the professional outcome. In this paper, we propose an active learning method called “Let us join” which is intended to address the professional outcomes. These activities are necessary in today’s scenario of education as knowledge sharing and group learning among students is seldom observed. The process of conduction of the activity and attainment of POs are presented in this paper. A survey was carried out before and after the activity, looking at the social relationship, self-drive for initiating communication and ability to integrate the information and document the information. The activity assesses students and also grades them according to their contribution in the activity.

Assessment of Program Outcome by Open-Ended Experiment in Enzyme Technology Laboratory Course

The laboratory component of Enzyme Technology course is designed to provide hands-on experience to inculcate the experimental skill during isolation, purification and assay of various enzyme preparations. To improve the attainment of program outcomes through Enzyme Technology laboratory course, the experiments were categorized as demonstration, exercise, structured enquiry and open-ended experiment. The primary objective of open-ended experiments is to enable the student to design and conduct experiments, as well as to analyze and interpret data. The experiments were carried out in a group of two so that students shall develop the ability to function effectively as an individual and in a group with the capacity to be a leader as well as an effective team member. The present paper describes the design of the open-ended laboratory experiment and its impact on attainment of program outcome and the overall students’ learning experience. Student teams were given only with problem statement, keeping the study approaches open. The students were expected to carry-out literature review and design the experiment, with due consideration to resources and feasibility. All the students were graded by adopting evaluation rubrics. The evaluation results show significant improvement in the achievement of program outcomes. It is observed that design of experiments as open-ended one improves the learning experiences and nurtures the innovativeness and creativity among students.

Comprehensive Survey on Metal Artifact Reduction Methods in Computed Tomography Images

Over the past few years, medical imaging technology has significantly advanced. Today, medical imaging modalities have been designed with state-of-the-art technology to provide much better in-depth resolution, reduced artifacts, and improved contrast -to-noise ratio. However in many practical situations complete projection data is not acquired leading to incomplete data problem. When the data is incomplete, tomograms may blur, resolution degrades, noise increases and forms artifacts which is the most important factor in degrading the tomography image quality and eventually hinders diagnostic accuracy. Efficient strategies to address this problem and to improve the diagnostic acceptability of CT images are thus invaluable. This review work, presents comprehensive survey of techniques for minimization of streaking artifact due to metallic implant in CT images. Problematic issues and outlook for the future research are discussed too. The major goal of the paper is to provide a comprehensive reference source for the researchers involved in metal artifact reduction methods.