Voice Assistant for Hearing and Speech-Impaired Individuals Using Web framework

This study addresses challenges encountered in real-time hand sign language recognition for web applications, with the overarching goal of improving communication between sign language users and non-signers. Recent advancements in techniques such as Random Forest, Support Vector Machine (SVM), and Gradient Boosting Classifier show promise, yet implementation faces hurdles. Variability and complexity in hand gestures, computational demands for real-time processing, environmental factors like lighting variations, background clutter, and accommodating diverse user characteristics all pose significant challenges. The objective is to develop a robust system that effectively integrates ensemble learning methods and advanced hand tracking algorithms, complemented by audio feedback, to facilitate seamless interaction. By prioritising diverse gesture models and implementing adaptive techniques to mitigate environmental challenges, the system aims to achieve high accuracy and real-time recognition. Initial experimentation reveals promising model scores: Random Forest at 95.45%, SVM at 79.123%, and Gradient Boosting at 94.456%. Rigorous evaluations will be conducted to assess accuracy, speed, and user feedback comprehensively. Ultimately, this innovative technology endeavours to bridge communication gaps and empower individuals with speech and hearing impairments, thus fostering inclusivity in society.