MUMBAI, India, March 14 -- Intellectual Property India has published a patent application (202441068755 A) filed by Indian Institute Of Technology, Tirupati, Andhra Pradesh, on Sept. 11, 2024, for 'real-time video-stream based dynamically configurable person specific models for group attendance.'
Inventor(s) include Kalidas Yeturu.
The application for the patent was published on March 13, under issue no. 11/2026.
According to the abstract released by the Intellectual Property India: "The present invention relates to a real-time video-stream based system for dynamically configurable person-specific models aimed at group attendance tracking, utilizing hand-held camera devices like smartphones. As a user walks through a classroom, a continuous video stream captures successive image frames, marking attendance for recognized faces. The system enhances facial sub-regions through super resolution before identifying individuals. It features a novel model scheme allowing dynamic configuration of person model pools for different groups. This invention also includes innovative computer vision techniques for miniature face recognition models. Video streams are captured at specified frame rates and sent to a backend server, which performs face region segmentation and recognition. Individual person models can be retrained to adapt to changes in facial features. The training algorithm involves synthetic group photo generation, face region extraction, and face embedding generation, facilitating robust biometric identification."
Disclaimer: Curated by HT Syndication.
