Intellectual Property India Publishes Patent Application for Enhanced Deepfake Image Detection Using Texture-Aware Preprocessing With Mesonet Architecture

MUMBAI, India, Oct. 26 -- Intellectual Property India has published a patent application (202611114870 A) filed by Mangalmay Institute Of Management & Technology on September 25, 2026, for Enhanced Deepfake Image Detection Using Texture-Aware Preprocessing With Mesonet Architecture.

Inventors include Ms. Lucky Rajpoot; Ms. Ambica Yadav; Ms. Mala Yadav; Mrs. Kimmi Gupta; Ms. Meha Mathur; Mr. Dileep Kumar G; Ms. Deepti Gautam; Dr. Sneh Kumari; and Mrs. Deepika Saxena.

The application for the patent was published on October 23, 2026, under issue no. 43/2026.

Abstract: The present invention provides an enhanced deepfake image detection system (100) and a method (200) thereof. The system (100) comprises an image acquisition module (110), a face detection and alignment module (120) generating an aligned facial crop, and a texture-aware preprocessing module (130) in which a local binary pattern transformation unit (132), a high-frequency residual extraction unit (134) and an adaptive contrast enhancement unit (136) operate concurrently upon the crop, a channel fusion unit (138) stacking the resulting maps into a multi-channel texture tensor. A MesoNet based classification module (150) having mesoscopic convolutional blocks (152), an inception style block (154), a classification head (156) and a transfer learning unit (158) classifies the tensor. A data augmentation module (140), a training module (160), an evaluation module (170) and a decision module (180) executed upon a computing platform (190) yield a calibrated forensic verdict with high accuracy and low computational cos

Disclaimer: Curated by HT Syndication.