Coronary artery diseases are one of the high-risk diseases, which occur due to the insufficient blood supply to the heart. The different types of plaques formed inside the artery leads to the blockage of the blood stream. Understanding the type of plaques along with the detection and classification of plaques supports in reducing the mortality of patients. The objective of this study is to present a novel clustering method of plaque segmentation followed by wavelet transform based feature extraction. The extracted features of all different kinds of calcified and sub calcified plaques are applied to first train and test three machine learning classifiers including support vector machine, random forest and decision tree classifiers. The bootstrap ensemble classifier then decides the best classification result through a voting method of three classifiers. A training dataset including 64 normal CTA images and 73 abnormal CTA images is used, while a testing dataset consists of 111 normal CTA images and 103 abnormal CTA images. The evaluation metrics shows better classification rate and accuracy of 97.7%. The Sensitivity and Specificity rates are 97.8% and 97.5%, respectively. As a result, our study results demonstrate the feasibility and advantages of developing and applying this new image processing and machine learning scheme to assist coronary artery plaque detection and classification.
Cardiac arrest in human arises owing to blood vessel diseases or heart defects. Blood vessel diseases result due to the blockage of blood in the heart vessels, which leads to pain in the heart. Heart defects occur because of damage in the cardiac muscles indicated by abnormal heart rhythms. Cardiovascular diseases cause mortality which could be avoided through the earlier detection of cardiovascular diseases. The major cause for cardiovascular diseases is cholesterol deposition inside the artery walls which later forms plaques that block the blood flow. Until now, plaques have been detected through medical imaging only after the heart attack. The plaques are blasted through angioplasty or reduced with medicine. Classification of the plaques before treatment, leads to effective medication based on the type of plaque. The sub classification of the plaque types such as rupture-prone plaque, ruptured plaque with sub occlusive thrombus, erosion-prone plaque, calcified nodule and non-plaque has been segmented and identified. In this paper, we propose a novel Spatial Fuzzy Propensity Score Matching (SFPSM) method to classify the plaques. The SFPSM method consists of clustering, ranking the cluster and region-based pixel wise analysis. Pixel analysis inspects specific regions of sub pixel points and calibrates the plaque. From the experimental results, the classification of plaque based on the 50-image data set has exhibited accuracy of 85% after validation. The plaque accuracy of classification provides the standard digital number values for the sub classification of plaques.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
customersupport@researchsolutions.com
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.