Design and Evaluation of a Model for Detecting and Identifying the Optimal Vein Using Artificial Intelligence Algorithms

Abstract

This paper aims to design and develop an integrated system based on Artificial Intelligence and computer vision technologies for the highly accurate detection of superficial veins, with the goal of supporting medical practices and reducing errors during injection and blood withdrawal procedures. The study followed an experimental–applied approach, divided into several stages, starting with data collection from publicly available infrared vein image databases, followed by training and testing models using three different algorithms ( YOLOv11, U-Net, and ORB ) and then analyzing the results to select the most efficient model. The test results showed that the YOLOv11 algorithm outperformed the others, achieving high accuracy in vein detection and making it the optimal choice for implementation in the final system. An interactive interface was developed using Python and the Tkinter library to enable users to upload images, run algorithms, and display results conveniently and efficiently. A field evaluation was conducted at Sabratha Teaching Hospital using a questionnaire to measure the system’s effectiveness and acceptance among healthcare practitioners, and the results demonstrated a high level of satisfaction. This study highlights the importance of integrating Artificial Intelligence technologies into medical applications to improve the quality of healthcare services and reduce challenges associated with vein identification procedures.