Empowering Hearing Impaired Communication Through Gesture Interpretation In Indian Sign Language Using Deep Learning

Indian Sign Language (ISL) is pivotal for the deaf community in India, aiming to surmount communication barriers and promote inclusivity. However, conveying technical nuances through ISL faces impediments rooted in cultural, linguistic, and resource constraints. To surmount these challenges, a standardized system was devised, leveraging specialized dictionaries. This initiative primarily targets enhancing communication within the Deaf community, particularly in technical domains. Employing image processing techniques for input preprocessing and harnessing the prowess of an EfficientNet-B0 Convolutional Neural Network, driven by deep learning algorithms, this system generates precise outputs. Supported by a vast dataset comprising digits, characters, and expressive gestures— meticulously categorized into numbers (0-9) and letters (A-Z), each containing 1,200 images—the repository also encompasses 22 descriptive words complemented by videos, enriching gesture recognition capabilities. This exhaustive dataset substantially bolsters real-time accuracy and precision in deciphering technicalities within ISL. This innovation not only aids inclusivity advocates and organizations but also facilitates enhanced sign language learning across diverse domains. Notably, the EfficientNet-B0 CNN-based methodology boasts an impressive 99.8% accuracy and efficiency, contributing profoundly to fostering a more interconnected, inclusive society where seamless communication prevails across all realms. Remarkably surpassing existing systems, this innovation showcases an exceptional 99.8% accuracy. These advancements cater to inclusivity advocates and organizations, facilitating enriched sign language learning across diverse domains. Ultimately, this system contributes to fostering a more connected, inclusive society by enabling seamless communication across all realms, showcasing unparalleled technical proficiency with its 99.8% accuracy rate.