MUMBAI, India, July 7 -- Intellectual Property India has published a patent application (202521086255 A) filed by Pandit Deendayal Energy University on September 11, 2025, for Portable Offline Translator For Sign Language And Speech Using Deep Learning.
Inventors include Deep Mathukiya; Jenil Patel; Darshil Padmani; Ansh Sardhara; Dr. Debabrata Swain; Dr. Kaibalya Prasad Panda; Dr. Davinder Paul Singh; Dr. Manish Kumar; Kashyap Vekariya; Mansi Parmar; and Darshan Rajani.
The application for the patent was published on July 03, 2026, under issue no. 27/2026.
Abstract: The titled invention discloses a system and method for real-time bidirectional translation between sign language and speech, the system comprising a Jetson Nano processing module(1), a camera (2) for gesture capture, a speaker (3), a microphone(4) , a touch screen based LCD screen (5) for mode selection and sign rendering, to activate gesture or speech input modes; wherein the Jetson Nano (1) is responsible for processing either image frames or audio input through deep learning models to perform sign-to-speech or speech-to-sign conversion. In sign mode, a gesture performed by a user is captured by the camera (2) and processed into text and speech. In speech mode, spoken words are converted to text and rendered into visual sign representations on the screen. The system operates entirely offline, ensuring real-time functionality without reliance on internet connectivity. The present invention is a cost- effective and inclusive solution accessible to the general public, supporting communication between hearing-impaired individuals and non-signers. By enabling dual mode translation in a portable embedded system, this invention promotes accessibility and seamless communication for all users.
Disclaimer: Curated by HT Syndication.
