MUMBAI, India, Jan. 23 -- Intellectual Property India has published a patent application (202521124761 A) filed by Shri Shankracharya Institute Llp, Durg, Chattisgarh, on Dec. 10, 2025, for 'system and method for providing ai powered reverse engineering of video with one-click reuse.'
Inventor(s) include Nishant Tripathi; Jatin Kumar Sahu; Vaibhav Raj Sahu; Dr. D. S. Kshatri; Dr. Shubhra Mishra; Dr. Anita Singh; and Dr. Sanjay Tiwari.
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: "A system (100) and method for providing AI powered reverse engineering of video with one-click reuse is disclosed. The controller (106) provides detect and reconstruct authentic video content from a deepfake video by performing steps of receiving an input video suspected of containing synthetic or manipulated content. The steps further include analyzing the video using a trained deep neural network configured to detect deepfake characteristics based on frame-level artifacts, temporal inconsistencies, and biometric mismatches. The steps further include determining whether the input video comprises at least one manipulated segment. The steps further include extracting latent feature representations of the manipulated segment. The steps further include generating a reconstructed authentic representation of the manipulated segment by performing generative inversion into a pre-trained authenticity model trained on original biometric or contextual features. The steps further include outputting a restored video comprising a replacement of the manipulated segment with the reconstructed authentic representation."
Disclaimer: Curated by HT Syndication.
