MUMBAI, India, Oct. 5 -- Intellectual Property India has published a patent application (202641115762 A) filed by Srm Institute Of Science And Technology, Ramapuram Campus; and Easwari Engineering College on September 28, 2026, for Dynamic Double Handed Indian Sign Language Recognition For Hearing And Speech Impaired People Using Hybrid Deep Learning Model.
Inventors include Dr. G. K. Vaidhya; Dr. V. Jananee; and Dr. A. Senthilselvi.
The application for the patent was published on October 02, 2026, under issue no. 40/2026.
Abstract: Abstract Human interaction is fundamentally enabled by communication, yet individuals with hearing impairments frequently encounter communication challenges due to limited access to trained interpreters. In India, Indian Sign Language (ISL) acts as a fundamental mode of communication, but its dynamic and double-handed gestures present significant challenges for automated recognition systems. These gestures involve complex coordination, temporal dependencies, and variations in speed, along with real-world issues such as lighting changes, occlusions, and background noise. To address the identified problems, this patent proposes hybrid CNN– Transformer architecture for dynamic double-handed ISL recognition. Spatial information, including hand posture, orientation, and inter-hand relationships, is generated from frame-wise visual data through Convolutional Neural Networks (CNNs). The extracted feature vectors are then processed by a Transformer module, which captures temporal relationships across sequences using self-attention mechanisms. This integration enables effective modeling of both spatial and temporal characteristics of gestures. The model is evaluated on both a public benchmark dataset and a custom ISL dataset, demonstrating strong generalization and efficiency. The proposed system has practical applications in education, healthcare, and assistive technologies, contributing to equal access to information and assists individuals with hearing impairments in overcoming communication barriers.
Disclaimer: Curated by HT Syndication.
