MUMBAI, India, Feb. 27 -- Intellectual Property India has published a patent application (202641017662 A) filed by Sri Krishna College Of Engineering And Technology, Coimbatore, Tamil Nadu, on Feb. 17, for 'ai-based system for generating text from visual lip movements in video streams.'
Inventor(s) include Dr. Maheswaran C P.
The application for the patent was published on Feb. 27, under issue no. 09/2026.
According to the abstract released by the Intellectual Property India: "The present invention discloses an artificial intelligence-based visual speech recognition system for generating text from lip movements captured in video streams without reliance on audio signals. The system comprises an input video source (100) configured to capture facial video data, a lip region extraction module (110) for isolating lip movements, and a preprocessing module (120) for normalizing video frames. A three-dimensional convolutional neural network feature extraction module (130) extracts spatio-temporal lip features, which are processed by a bi-directional recurrent neural network module (140) to model temporal speech patterns. A connectionist temporal classification decoding module (150) converts the learned features into character sequences, optionally refined by a language correction module (160). An adaptive speaker calibration module (170) enables dynamic personalization for different speakers. The final textual output (180) provides accurate transcription of spoken content from visual data, enabling silent communication, assistive technology, and speech recognition in noisy environments."
Disclaimer: Curated by HT Syndication.
