Loading...

Determination of the Degree of Occlusion of Coronary Arteries by Processing Angiography Images

Ghalehnovi, Mahboobeh | 2014

879 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 45964 (05)
  4. University: Sharif University of Technology
  5. Department: Electrical Engineering
  6. Advisor(s): Zahedi, Edmond; Fatemizade, Emadeddin
  7. Abstract:
  8. Cardiovascular disease is considered as the most important cause of death in the world. The coronary vessels, with three main arteries, has quite important. Coronary angiography is still the most common modality for physicians to assess the severity of vessel narrowing or stenosis during percutaneous coronary intervention procedure. For this procedure, a thin, flexible tube called a catheter is put into a blood vessel in your arm, groin (upper thigh), or neck. The tube is threaded into your coronary arteries, and the dye is released into your bloodstream. X-ray pictures are taken while the dye is flowing through the coronary arteries. Physicians evaluate angiographic images visually with using the experience, or the Gensini valuation criteria as the most widely used method for scoring angiographic images. To prevent human errors such as not having sufficient expertise or user fatigue, eye identifying and those cases, we need a method that segmentation of arteries be done automatically and the degree of occlusion is determined. In this study, 50 sets of data has been used, were recorded at Tehran Heart Hospital. Physicians, has classified the data to the six classes in terms of the degree of occlusion: class Zero (0% stenosis), class I (stenosis less than 50%), class II (50-70% stenosis), class III (70 - 90% stenosis), class IV (90-99% stenosis) and class V (100% stenosis). In this study, the method used to determine of the degree of occlusion of coronary artery, consists of several steps: first, to improve images, noise removed with BM3D method and then segmentation of coronary arteries be done with active contour models and level set approach. Continue to identify and determine the degree of occlusion, first the central axis of the arterial tree was extracted by thinning the binary image with morphology operators and second with iterative algorithm of finding the endpoints of each branch and then remove them from the central axis, the central axis of each branch was separated from each other. After sorting the branches of the central axis, the diameter of each branch is determined by finding two points around each point of each branch and finding the minimum distance. In the last stage with thresholding and windowing around each pixel on every branch, location and percent of degree of occlusion of arteries is determined due to normal vessels surrounding atherosclerotic vessels. The proposed method has been implemented for three images from each data set and the degree of occlusionis estimated forfour arteries: left main coronary artery, circumflex artery, anterior descending artery and right coronary artery. The results of the accuracy percent of proposed method compared with physicians eye diagnosis is 90% for class Zero, 90% for class I, 98% for class II, 99% for class III, 99% for class IV and 98% for class V. However, due to the diagnosis of stenosis less than 10% is difficult for physicians then in mentioned results, the amount of stenosis less than 10% have been achieved with the proposed algorithm have been placed into the classZero
  9. Keywords:
  10. Coronary Arteries ; Active Contour Model ; Occlusion Detection ; Vascular Tree Extraction ; Angiography Image Segmentation ; Fuzzy Level Set

 Digital Object List

 Bookmark

No TOC