MUMBAI, India, Aug. 12 -- Intellectual Property India has published a patent application (202621075615 A) filed by Ramdeobaba University, Nagpur; Shri Ramdeobaba College Of Engineering And Management, Nagpur; Thakkar, Avni S.; Hardas, Bhalchandra M.; and Patil, Prashil on June 18, 2026, for A System For Real-Time Deepfake Detection And Prevention During A Live Digital Communication Session.
Inventors include Thakkar, Avni S.; Hardas, Bhalchandra M.; and Patil, Prashil.
The application for the patent was published on August 07, 2026, under issue no. 32/2026.
Abstract: A System for Real-Time Deepfake Detection and Prevention During a Live Digital Communication Session The present invention relates to a system (100) for real- time deepfake detection and prevention during a live digital communication session. The system (100) includes a user device (120), an input acquisition module (140), hardware sensor correlation module (160), biological liveness module (180), behavioural biometric profiling module (200), virtual camera and stream detection module (220), an audio authenticity module (240), a network cadence verification module (260), zero-trust decision engine (280), and an action gateway (300). The system (100) continuously acquires video, audio, device motion sensor data, and network transmission metadata. The system detects micro-jitters correlated with visual micro- movement, photo pupillary response to display luminosity, person-specific behaviour is profiled, and network cadence is analysed. The zero- trust decision engine (280) generates a liveness score, authenticity ranking, and deepfake probability score to determine risk conditions, and the action gateway (300) issues warnings, requests secondary verification, soft-locks sessions, denies access, or terminates sessions.
Disclaimer: Curated by HT Syndication.
