MUMBAI, India, Jan. 23 -- Intellectual Property India has published a patent application (202511125240 A) filed by Uttaranchal University, Dehradun, Uttarakhand, on Dec. 11, 2025, for 'an automated image plagiarism detection using distinct deep learning techniques modified cnn, siamese network, and vision transformer architectures.'
Inventor(s) include Anjali Naudiyal; Kapil Joshi; and Atul Bhatt.
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: "The invention discloses an automated system for detecting image plagiarism using a hybrid deep learning framework that integrates Modified CNN, Siamese Networks, and Vision Transformer architectures. Traditional methods relying on pixel-level comparison and handcrafted features fail under transformations such as cropping, rotation, scaling, brightness adjustment, watermarking, and semantic modifications. The proposed system overcomes these limitations by combining CNN for detailed feature extraction, Siamese Networks for pairwise similarity detection, and Vision Transformers for contextual pattern recognition. Simultaneously, OCR modules extract textual content, enabling multimodal analysis of both visual and textual features. The system compares extracted features against large reference datasets, computes similarity scores, and evaluates authenticity by analyzing metadata and watermarks. A plagiarism report is generated, quantifying similarity percentages, highlighting plagiarized regions, and identifying sources of copied content. The invention provides a scalable, explainable, and real-time solution suitable for academic, industrial, and professional applications requiring reliable image originality verification."
Disclaimer: Curated by HT Syndication.
