MUMBAI, India, July 7 -- Intellectual Property India has published a patent application (202641076540 A) filed by Sasi Institute Of Technology & Engineering; Yasoda Khumbham; Rambabu Pasumarthy; M Anantha Lakshmi; and M V V A L Sunitha on June 20, 2026, for Deep Learning Based American Sign Language Recognition System With Mediapipe And Bi-Lstm.
Inventors include Yasoda Khumbham; Rambabu Pasumarthy; M Anantha Lakshmi; and M V V A L Sunitha.
The application for the patent was published on July 03, 2026, under issue no. 27/2026.
Abstract: Communication barriers faced by people with hearing and speech impairments drive significant research interest in developing assistive technologies for sign language recognition. This review focuses on real-time sign language to text and speech conversion with multilingual translation, a domain at the intersection of Artificial Intelligence, Deep Learning and HumanComputer Interaction. The main objective is to analyse and compare existing methodologies, classify approaches, and identify gaps to guide future research in creating better communication systems. This review adopts a scoping review and methodology, covering studies published after 2015 that utilise computer vision and deep learning techniques for gesture recognition and multilingual translation. Key findings highlight of vision-based deep learning methods, particularly Convolutional Neural Networks (CNNs) and media pipe with Neural Network models for dynamic gesture recognition. However, gaps remain in integrating multilingual and multilingual translations, training models with dynamic datasets, and making sure of low-latency real-time performance. This review contributes anew group of sign language recognition systems, comparative analysis of existing methods and a framework combining gesture recognition, text generation and speech generation into a unified pipeline
Disclaimer: Curated by HT Syndication.
