Computational Methods for Detection and Recognition of Coronary Artery Stenosis in Angiogram Images

Title

Computational Methods for Detection and Recognition of Coronary Artery Stenosis in Angiogram Images

Creator

Kavipriya, K

Contributor

Hiremath, Manjunatha

Description

Coronary Artery Disease (CAD) is caused by stenosis of the coronary artery's lumen. This heart disease is one of the reasons for the highest mortality worldwide. This illness manifests as stenosis or plaque in the coronary arteries and causes atherosclerosis. It damages or clogs the heart arteries, causing a lack of blood flow to the heart muscles and leading to a heart attack. There are different medical modalities to diagnose the heart artery disease. A standard method used by the cardiologist to diagnose the severity of this disease is coronary angiography. An X-ray machine is used to capture the angiogram image at various angles during cardiac catheterization. Experts examine the data and offers different opinions. owever, most of the angiogram videos consist of unclear images with artifacts, and because of the complex structure of the arteries, medical experts fail to get accurate information about the damages and blockages in arteries. Based on the cardiologist's suggestions, a computational model is proposed as a secondary method to detect and recognize the stenosis level from the coronary angiogram images. The proposed model is Coronary Artery Stenosis Detection Using Digital Image Processing (CASDDIP). The proposed research model/framework can identify the stenosis in the cardiogram image with good accuracy of 98.06% precision. This proposed research experimentation can be compared with existing literature methods which outperforms compared to other methods using real time dataset. A dataset, such as angiogram videos and images of patients under varying age groups, is used to train the model. These videos are acquired from the healthcare center with due consent. The proposed CASDDIP model consists of four modules: Keyframe extraction and preprocessing Coronary Artery Segmentation Feature extraction and stenosis detection Initially, a novel keyframe extraction method is proposed to find the keyframe from the angiogram video. Followed by a hybrid segmentation method is presented in this research to extract the coronary artery region from the image. Further a method is proposed to detect the stenosis by extracting and fusing different features. Detected stenosis is categorized using the proposed stenosis level classification method. This CASDDIP model is a supporting tool to help the cardiologist during diagnosis.

Source

Author's Submission

Date

2024-01-01

Publisher

Christ(Deemed to be University)

Subject

Computer Science

Rights

Open Access

Relation

61000344

Format

PDF

Language

English

Type

PhD

Identifier

http://hdl.handle.net/10603/575076