TY  - BOOK
AU  - De Leon,Chrzel Kyara C.; Lacsamana,Christian G.; Ramos,Jarius Rian I.
TI  - FAR (First aid response): Augmented reality mobile application for real-time wound assessment and treatment instructions using convolutional neural network, roboflow and cumulative histogram equalizer
AV  - T58.64 D45 2025 
PY  - 2025///
CY  - .
PB  - .
KW  - academic writing
N1  - ABSTRACT: This study developed FAR (First Aid Responder), an augmented reality (AR)-based mobile application that delivers real-time wound assessment and first aid instructions using Convolutional Neural Networks (CNN), Roboflow, and Cumulative Histogram Equalization (CHE). Aimed at addressing the lack of accessible first aid tools in emergency situations especially for individuals without medical training. FAR focuses on three common wound types: bruises, abrasions, and burns. Through an intuitive AR interface, the app provides step-by-step visual guidance for treatment. To ensure accurate wound detection in non-clinical settings. CNNs were used for image classification. The models were trained using Roboflow, and image clarity was enhanced using CHE, improving the accuracy of wound recognition. The primary objective was to develop an intelligent AR-based system capable of identifying wounds and recommending treatment in real time. Specific objectives included integrating live image processing, automating symptom analysis, and enhancing wound visibility through contrast improvement. The application was developed for the Android platform and tested using a custom dataset trained in Roboflow. Performance results showed the CNN model reached a wound detection accuracy of up to 86%, while CHE enhanced image contrast by approximately 75%, making wounds easier to classify. Roboflow metrics, including precision, recall, and mean average precision (mAP), confirmed the model’s reliability across various image conditions. In conclusion, FAR demonstrates the promising integration of AR and AI for practical, real-time wound assessment. It serves as an accessible, educational, and potentially life-saving tool. Future enhancements may include severity analysis, hospital locator integration, and dataset expansion for broader wound classification; F
ER  -