Leveraging Advanced AI for Real-Time Sign Language Detection and Translation: A Pioneering Approach to Bridging Communicative Gaps and Promoting Societal Inclusivity

Embarking on a journey through the complex realm of advanced artificial intelligence (AI) methodologies, this study is driven by the deep urgency to eliminate the communication barriers that individuals with auditory, vocal, or visual impairments face on a daily basis. Using the power of cutting-edge object detection models: YoLOv8, YoLOv9, YoLOv10, and Detectron2, our study showcases exceptional precision and real-time responsiveness in sign language detection and translation, even within the constraints of a limited dataset that includes seven specific gestures. The investigation underscores the pressing need to expand the data set and integrate the Florence2 model to further enhance detection accuracy and robustness of the system. The proposed system, implemented on a Raspberry Pi 4 platform and enriched with real-time data capture through a camera and audio output through a microphone, is envisioned as a groundbreaking wearable device, such as smart sunglasses. This innovative solution, which converts visual gestures into spoken language, is poised to revolutionize communication accessibility in Bangla-speaking regions, offering a pioneering approach to bridge communication gaps. The study paves the way for future advancements, including the incorporation of Bangla Sign Language, positioning this work at the forefront of AI-driven assistive technology aimed at fostering social inclusion.