Study of heart disease patient identification and monitoring methods based on the YOLOv8 HEART

Cardiovascular disease is the leading cause of death worldwide, resulting 2es2w2qto more than 500,000 deaths, so cardiovascular disease behavioral monitoring is crucial. Sudden symptoms include heartache, angina pectoris, and shortness of breath. The YOLO algorithm performs well in the field of target detection and is widely used in many medical object detection tasks. The YOLOv8_HEART model presented in this paper is based on YOLOv8 for real-time monitoring of a medical care system of cardiovascular patients during hospitalization or at home. Some abnormal movements of patients at the onset of the disease can be used to identify the discomfort. The paper has collected a large number of data sets to identify the above movements. Secondly, considering the real-time nature of the sudden disease and emphasizing the rate of the model, YOLOv8_HEART improves the network structure and emphasizes behavior-specific detection in various visual contexts. The sensitivity of the model to abnormal behavior was improved by customized preprocessing and architecture adjustments. The results show that YOLOv8_HEART performs well on the current medical care data set, with a mAP score of 0.848 and an inference speed of 151.515fps.