MUMBAI, India, Sept. 28 -- Intellectual Property India has published a patent application (202641113123 A) filed by Dr. M. G. R. Educational & Research Institute on September 21, 2026, for A System And A Method For Sensor-Fusion Gesture-To-Language Translation And Speech Generation Communication.
Inventors include K. Sujatha; B. Swapna; A. Jegadeeswari; V. V. Kannan Suresh; and B. Bhavana Priya.
The application for the patent was published on September 25, 2026, under issue no. 39/2026.
Abstract: Disclosed is a system (100) for sensor-fusion based gesture-to-language translation of physical gestures into target-language text and synthesized speech. The system (100) comprises a multimodal sensor array (102), an embedded processing unit (104), a preprocessing module (106), a temporal segmentation module (108), a gesture recognition module (110), a neural machine translation module (112), a text-to-speech module (114), an output interface (116), and a personalization module (118). The multimodal sensor array (102) captures finger-bend, inertial, and image information. The temporal segmentation module (108) identifies gesture onset and offset during continuous signing. The gesture recognition module (110) performs multimodal sensor fusion to generate linguistic tokens and resolve gesture ambiguities. The neural machine translation module (112) converts the linguistic tokens into target-language text, while the text-to-speech module (114) generates synthesized speech. The personalization module (118) provides few-shot user adaptation, enabling continuous, adaptive, privacy-preserving, and substantially on-device gesture translation with reduced external computing dependence. Fig. 1 is the reference figure.
Disclaimer: Curated by HT Syndication.
