BRAZILIAN SIGN LANGUAGE RECOGNITION AND TRANSLATION APPLICATION

Abstract

This paper presents the development of the LibrasCam mobile application, which enables real-time recognition and translation of Brazilian Sign Language (Libras). Using computer vision algorithms and libraries such as OpenCV, MediaPipe, and TensorFlow, the system interprets gestures captured by the camera and converts them into text. The project was structured in stages including requirements gathering, AI model development, implementation in Flutter, and remote server integration. The results indicated high accuracy and adequate response time, demonstrating the solution's potential to promote accessibility and social inclusion