Capsure: A Novel Application of CNNs, OpenAl's Language Model, and Google Text-to-Speech for Reliable Drug Recognition in the Philippines

This study presents the development and evaluation of CapSure, a web-based application designed to facilitate pharmaceutical product identification for individuals in the low-vision community. Using advanced computer vision technology, CapSure accurately classifies pharmaceutical products, providing relevant information to users, thus enhancing their independence and quality of life. The study involved a two-fold evaluation process: a comprehensive performance assessment of the underlying image classification model, and a user-focused survey based on ISO/IEC 25010 software quality standard. The image classification model exhibited an impressive overall accuracy of 98.65%, demonstrating its robustness and effectiveness. A usability survey, administered to 15 participants from the low-vision community, reported high scores across all sub-characteristics of the ISO 25010 standard, validating CapSure's effectiveness and user-friendliness. Future enhancements may include the integration of voice commands and the expansion of the pharmaceutical product database.