Empowering Individuals with Disabilities: Gesture-to-Voice Technology for Enhanced Communication

Abstract

Gesture-to-Voice technology offers a promising solution for improving communication between individuals with speech or hearing impairments and the wider community. This paper focuses on camera-based recognition approaches, which rely on widely available devices such as webcams and smartphones. A review of recent methods highlights the transition from CNNs for static gestures to spatiotemporal models and transformers for dynamic recognition, with lightweight networks enabling real-time applications. Although accuracy has improved, challenges remain in robustness to environmental noise, signer independence, and multilingual support. By analyzing existing research, datasets, and tools, this study identifies gaps and provides recommendations for developing affordable, offline, and culturally adaptive Gesture-to-Voice systems, particularly in the Indian context.