MUMBAI, India, July 7 -- Intellectual Property India has published a patent application (202641077272 A) filed by C M R Institute Of Technology; Mr. Arvind R; Ms. Rajeshwari R; Ms. Lynsha Helena Pratheeba; Ms. Smruthi Nair; Manju V N; and . Mr. Abhilash R on June 23, 2026, for Edge Ai–iot System And Method For Real-Time Multimodal Deepfake Detection With Blockchain Enabled Forensic Evidence Certification.
Inventors include Manju V N; Ms. Smruthi Nair; Ms. Lynsha Helena Pratheeba; Ms. Rajeshwari R; Mr. Arvind R; and . Mr. Abhilash R.
The application for the patent was published on July 03, 2026, under issue no. 27/2026.
Abstract: An edge AI–IoT system and method are disclosed for real-time multimodal deepfake detection with blockchain-enabled forensic evidence certification. The system comprises one or more IoT edge nodes equipped with embedded AI processors (for example NVIDIA Jetson Nano, Raspberry Pi integrated with AI accelerators, or equivalent edge inference hardware) capable of executing optimized deep learning models locally without reliance on remote cloud infrastructure. Real-time multimodal analysis is implemented directly on the edge device using quantized cross-modal attention models with sub-second inference latency. The multimodal framework analyzes multiple data modalities including visual facial features extracted from video frames, acoustic speech characteristics derived from audio streams, and associated media metadata and audio–visual synchronization patterns to detect manipulated or synthetically generated multimedia content. Upon detection of suspected deepfake media, the edge node autonomously generates a forensic evidence package including a cryptographic hash of the detected media segment, detection confidence score, device identifier, model provenance information, and a hardware-secured digital signature. The evidence package is submitted to a permissioned distributed blockchain network through a smart contract that immutably records the detection event, generates a verifiable forensic certificate compliant with digital evidence standards, and initiates an automated alert workflow for authorized stakeholders. The invention addresses limitations of existing systems including dependence on cloud-based analysis, unimodal detection approaches, lack of autonomous forensic evidence generation, and absence of legally verifiable certification mechanisms. The proposed system enables secure, privacy-preserving, and real-time verification of multimedia authenticity for applications including law enforcement, judicial proceedings, election integrity monitoring, broadcast media verification, and financial fraud prevention.
Disclaimer: Curated by HT Syndication.
