Abstract
Abdominal Aortic Aneurysm (AAA) is a disease that progresses without symptoms and has a high mortality rate. Abdominal ultrasonography is the initial diagnosis of choice for this disease, and the presence or absence of an aneurysm is determined by the length of the inter-pericardial distance of the vessels depicted on the ultrasound image. On the other hand, because ultrasound images are unclear, the accuracy of vessel diameter measurement depends on the skill of the reader. Therefore, for early detection and accurate diagnosis of diseases, a system that automatically analyzes blood vessels from images and supports reading and diagnosis is required for ultrasound systems. Therefore, in this study, we developed and evaluated a system that has functions for detecting vascular regions and measuring the distance between vascular adventitia using an image segmentation model and a model learned with YOLOv10 on ultrasound images of the abdominal aorta.
