AI-BASED SMARTHEAL SYSTEM: A VOICE AND VISION ENABLED FIRST-AID ASSISTANT FOR REAL-TIME INJURY DETECTION

Abstract

SmartHeal is a system implemented in MATLAB and designed to detect, identify, and analyze wounds in real time, in low-resource and emergency setting, with the optimization to optimality of the reporting. The system detects the type of wounds including abrasions, burns, lacerations, and ulcers with the help of artificial intelligence and deep learning and also determines the degree of the wound using texture, structure, color, and other features. The hybrid type of segmentation isolates the wound site to enable consistent measurements between different forms of light and equipment. SmartHeal aids sustained monitoring through neighbour epididymis through sequential image comparison to measure healing, identify infection threats, based on clinical indicators rule, including persistent redness, swelling and slowing of closure. Care guidance is also produced as a result of the system in order to guide the non-experts to act accordingly. SmartHeal is fully functional without the use of the internet and makes use of transfer learning, model optimization and MATLAB Runtime to deploy effectively. Its scalable architecture displays an economical approach to remote wound assessment, enhancing prompt responses and aid in medical decisions with less clinical knowledge.