Deep Learning Based Detection of Cardiovascular Diseases from Medical Images
Year: 2026
Type: conference-paper
Abstract: In medicine, images are the main factor in determining the patient’s condition. In particular, medical images are important in the diagnosis of heart diseases, and they are also relevant, and the types of medical images are very different (magnetic resonance imaging, computed tomography, X-ray imaging, and other similar things). Like images, diseases are also different, especially early detection of changes in coronary arteries, which is considered cardiovascular disease, is important for effective treatment of the disease. Manual analysis of angiographic images used in the diagnosis of this disease is strongly dependent on the experience of the specialist and time factors. Therefore, by analyzing the work done so far, we proposed a deep learning approach that automatically detects vascular structures and perfusion areas from coronary artery images. The method presented in this work performs a twostage segmentation process. The nnU-Net architecture is presented as the first stage to segment arterial vessels, and in the second stage, a U-Net-type model is used to detect perfusion areas. The open-source ARCADE dataset was used in the study. Experimental results show that our proposed approach allows for efficient identification of syntenic regions.
Institution: Tashkent University of Information Technology
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W7202293408