MUMBAI, India, Aug. 10 -- Intellectual Property India has published a patent application (202521133907 A) filed by Sandip Institute Of Technology And Research Centre; Sandip Institute Of Engineering And Management; Sandip Polytechnic; Sandip Institute Of Pharmaceutical Sciences; Sandip University Nashik; and Sandip University Sijjoul on December 31, 2025, for A Deepfake Classification System For Human Facial Images And Videos.
Inventors include Bendale, Archana Vedant; Bage, Dipak D.; Rumale, Aniruddha S.; Mogare, Savita B.; Patil, Parag B.; Patil, Rupali R.; Devhare, Chaitanya; and Kapadnis, Vaibhavi.
The application for the patent was published on July 31, 2026, under issue no. 31/2026.
Abstract: The present invention relates to a deepfake classification system for human facial images and videos. The deep learning-based system for the automated classification of Deepfake human faces in images and videos. The invention addresses the growing challenges posed by synthetic media by utilizing advanced convolutional neural networks (CNNs) and feature extraction techniques to detect subtle artifacts, distortions, and inconsistencies introduced during Deepfake generation. The system processes input data through a sequence of steps including preprocessing, facial region detection, feature extraction, and classification using a trained deep learning model. It incorporates data mining methods for early-stage analysis and employs transfer learning and data augmentation to enhance accuracy and model robustness. The invention further adapts to newly emerging Deepfake techniques through scalable and flexible design components. Comparative evaluation demonstrates improved precision and reliability over conventional detection approaches.
Disclaimer: Curated by HT Syndication.
