Wound Detection and Analysis System

Abstract

Wound organization is a vital characteristic of healthcare that stresses fortunate and culminate organization to guarantee fitting treatment and decrease complications. This paper presents Wound Detox, a machine learning-based application created to classify wounds into five categories: Scraped spot, Bruise, Cut, Puncture, and Slash. These are gathered into three seriousness levels: low (Scraped area, Bruise), medium (Cut, Puncture), and high (Gash). The framework utilizes a Convolutional Neural Network (CNN) prepared on an assorted dataset of wound pictures for solid picture classification. Users can transfer wound pictures by means of a portable application built with React Native, which communicates with a Java-based backend. Upon classification, the app gives first-aid direction for low and medium-level wounds and appears adjacent clinics for medium cases. For high-severity wounds, the system automatically initiates an emergency call to the nearest ambulance service. The backend uses MySQL for secure data storage and efficient handling of user inputs and classification outcomes. User-friendly features such as registration, login, and image submission enhance usability. Wound Detox has been rigorously tested, demonstrating high accuracy and fast response times. By integrating AI with practical healthcare tools, the application offers accessible, real-time wound assessment and potentially life-saving support.