Presenting a robust algorithm for detecting coronary artery occlusion in multi-sectional CT-angio images and evaluating the accuracy of results with angiographic images

سال انتشار: 1399
نوع سند: مقاله کنفرانسی
زبان: انگلیسی
مشاهده: 426

فایل این مقاله در 8 صفحه با فرمت PDF و WORD قابل دریافت می باشد

استخراج به نرم افزارهای پژوهشی:

لینک ثابت به این مقاله:

شناسه ملی سند علمی:

TETSCONF05_013

تاریخ نمایه سازی: 25 دی 1399

چکیده مقاله:

Cardiovascular disease has increased dramatically in recent decades withpopulation growth and inactivity and unhealthy nutrition. It is estimated that by2030, more than 40% of the world's population will be infected with thedisease. According to world statistics in 2008, about 7249,000 people diedfrom these diseases, many of whom were victims under the age of 65.According to the statistics of the Ministry of Health of Iran in 1390, about 40%of deaths in Iran were due to cardiovascular disease. Heart disease includeshigh blood pressure, coronary artery disease, heart failure, heart attack andcongenital heart defects, which account for 54% of deaths from heart diseasedue to coronary artery disease. Early diagnosis of atherosclerosis is essential toprevent the emergence of serious risks in such patients with methods such asmedication and angioplasty. Because of some of the complications andproblems of invasive angiography, as well as some patients' fears and anxietiesabout having angiography, researchers have tried to find other methods that aremore acceptable and comfortable for patients. CT cardiac angiography is a newand diagnostic method for examining various arteries, including the coronaryartery, without the need for invasive cardiac angiography. In this study, usingautomatic processing (with the help of computer), CT angiographic images aredetected and classified with higher accuracy and precision of coronary arteryplaques. After that, the image was segmented by windowing the image to 60 *60 dimensions and applying Atso thresholding method, and then by extractinglocal and statistical features and Violet energy and high-order statistics, theeclipse site was classified by KNN, SVM, MLP and TREE. The results fordetermining coronary artery occulation with are classified with SVM classifiera polynomial kernel a 96% accuracy and a sensitivity of 90% and a detectioncapability of 100%, which indicates the potential of the proposed method toother articles.

نویسندگان

Mahsa Naghieh

Faculty of Science and Research University, Tehran, Iran

Leila Kalhor

Faculty of Islamic Azad University of Khomeini Shahr, Esfahan,Iran