AI-Integrated Diagnostic System for Varicose Veins using Deep Learning and Health Signal Inputs

Varicosities is a prevalent problem which belong to the adults category involved in prolonged standing and sitting conditions. Doppler ultrasonography and visual observation, are traditional techniques to diagnose varicose veins which demand the expertise of qualified medical professionals, and the high price of diagnostic technology presents a nearly insurmountable barrier to healthcare in developing countries. Thus this study proposed a novel idea of Intelligent Varicose Veins detection system. This system uses imaging and real-time health signals to provide affordable intelligent diagnostic solutions based on varicosity features and their relevant health signals. The deep learning model comprises of a hybrid framework combining YOLOv9 and EfficientNet- B0.The dataset consisted of about 2500 annotated leg images with above 94% accuracy and mean Average Precision of 89%. Different health parameters such as leg movement, pose and blood flow rate are measured through IMU and PPG/Doppler sensors. With efficient IoT device using ESP32-S3 Microcontroller, collected data is transmitted through a FastAPI and a React.js is used for vein diagnosis visualization and report generation. In Addition a cloud deployment model offers unparalleled flexibility, scalability, robust security, and seamless remote access.