چكيده به لاتين
Most deaths of heart disease are due to coronary heart disease. This condition is caused by the thickening of the arterial wall due to atherosclerotic plaque, which can cause narrowing or even obstruction of the arterial lumen. Currently, intravascular optical coherence tomography (IVOCT) is increasingly being used in clinics to diagnose coronary artery disease. Intravascular Optical Coherence Tomography (IVOCT) is a catheter-based invasive intravascular medical imaging system that utilizes near-infrared light to deliver high resolution cross-sectional images. Using automated procedures to segmented the coronary artery wall can save time and cost and can improve the accuracy of doctors' diagnosis of cardiovascular disease. In this thesis, a new automated method for the determination of the vessel wall is proposed and then the bifurcations are identified in the frames. Artery wall segmentation algorithm including pre-processing steps, determination of extremal regions of extremum levels (ERELs), selection of the best areas to be determined by classifying the boundaries into two boundary and non-boundary classes, and finally Mary Lumen and detected the lumen and media borders. And the bifurcation algorithm consists of the steps to determine the slits of each frame that can be caused by artifacts, plaques, and bifurcation, applying nonlinear regression to the lumen boundary value in the slit regions, and classifying these areas into bifurcation and non-bifurcation. We used real data to evaluate the algorithm, and the results of comparing the algorithm with manual segmentation show that the proposed algorithm, in addition to high accuracy in determining the lumen boundary (99% Dice criterion) and Hausdorff distance of
0.2mm indicates the stability of the algorithm. It has different conditions and also the algorithm has 93% accuracy for determining the bifurcation regions.
