MUMBAI, India, Nov. 2 -- Intellectual Property India has published a patent application (202641127422 A) filed by Rns Institute Of Technology; Preetham Heggade C R; Priya Javalagi; Chethan Gowda; Mithun Gowda K M; Keerthana K N; Priyanka A N; and Manjula Vasant Kiresur on October 26, 2026, for Ai-Based Sign Language Translation System.

Inventors include Manjula Vasant Kiresur; Priyanka A N; Chethan Gowda; Mithun Gowda K M; Keerthana K N; Priya Javalagi; and Preetham Heggade C R.

The application for the patent was published on October 30, 2026, under issue no. 44/2026.

Abstract: Abstract Sign language is the primary medium of communication for millions of deaf and mute individuals across the world. However, most hearing people are unaware of sign language, leading to significant communication barriers in everyday situations such as schools, hospitals, offices, and social environments. To address this problem, an AI-Based Sign Language Translation System was developed that uses a standard webcam to capture hand gestures and converts them into text and speech output in real time. The system integrates MediaPipe Hand Detection for accurate hand tracking and a Convolutional Neural Network (CNN) trained on 44 gesture classes for gesture classification. Each input frame is preprocessed using OpenCV through resizing and noise reduction before being passed to the detection and classification pipeline. A Stable Mode prediction filter ensures that only consistent predictions are displayed, eliminating flickering in the output. The recognized gesture is converted to text on a web-based dashboard and simultaneously synthesized into speech using the Pyttsx3 engine. A multilingual translation module supports output in Kannada, Hindi, and Telugu in addition to English. The system also supports a reverse text-to-sign pathway, allowing hearing users to type a word and view its corresponding sign gesture as an image or video, enabling genuine two-way communication. The system achieved a training accuracy of 94% and operates entirely on standard consumer hardware without requiring any GPU or specialized sensors. This work presents a reliable, accessible, and inclusive communication tool with strong potential to reduce the daily barriers faced by deaf and mute individuals.

Disclaimer: Curated by HT Syndication.