MUMBAI, India, Jan. 23 -- Intellectual Property India has published a patent application (202541117153 A) filed by Nandha Engineering College, Erode, Tamil Nadu, on Nov. 26, 2025, for 'hand2voice: enhancing accessibility with speech and sign language translation.'
Inventor(s) include K M Olimathi; Priyadarshikaa; S Ragavendiran; M Sathiyasudhan; and V Mythily.
The application for the patent was published on Jan. 23, under issue no. 04/2026.
According to the abstract released by the Intellectual Property India: "The present invention relates to a real-time gesture-to-speech communication system known as Hand2Voice, designed to translate human hand gestures into meaningful textual and auditory outputs. This invention integrates advanced computer vision techniques, time-series modeling, and lightweight deep learning architectures to create a non-invasive communication platform for individuals with speech or hearing impairments and for scenarios requiring silent or touch-free interaction. The system functions by continuously capturing video input usmg a standard camera module. A specialized landmark extraction engine identifies and tracks 21 anatomical key points on the user's hand, including fmgertip joints, knuckles, and wrist coordinates. These extracted landmarks are processed into structured sequences representing static and dynamic gestures with high accuracy. The invention utilizes a hybrid Convolutional Neural Network (CNN) and Long Short-Term Memol) (LSTM) model to interpret the landmark sequences. The CNN extracts spatial features from individual frames, while the LSTM analyzes temporal transitions necessary for recognizing dynamic gesture patterns. This FOmbined spatial-temporal analysis ensures robust classification performance even under challenging conditions such as varying illumination, background clutter, and partial occlusions. The system's lightweight design enables real-time performance on mid-range or embedded hardware. Once a gesture is accurately classified, the system automatically generates corresponding textual output, which is displayed on the user interface for immediate confirmation. An optional Text-to-Speech (ITS) module ponverts the recognized text into clear, natural-sounding audio,. enabling seamless gesture-to-voice Fommunication. The system incorporates reliability mechanisms that detect incomplete or ambiguous gestures and suppress incorrect predictions to maintain high trust and user satisfaction. Hand2Voice provides a completely automated, marker-less, and contact-free method for converting gestures into speech. With operational latency typically below 250 milliseconds, it is suitable for real-time conversations, assistive technology applications, remote communication platforms, and human-machine interaction systems. By integrating accuracy, efficiency, portability, and ease of deployment, this invention significantly advances gesture recognition technology and establishes a practical and scalable solution for bridging communication gaps between individuals and digital environments."
Disclaimer: Curated by HT Syndication.
