MUMBAI, India, Oct. 26 -- Intellectual Property India has published a patent application (202611100497 A) filed by Madan Mohan Malaviya University Of Technology on August 19, 2026, for A System And Method For Open-Set Cattle Identification.
Inventors include Nikita Tripathi; Dr. Satvik Vats; Dr. Shwet Ketu; Dr. Sneha Mishra; and Dr. Rakesh Kumar.
The application for the patent was published on October 23, 2026, under issue no. 43/2026.
Abstract: A computer-implemented prototype-based open-set livestock biometric identification system (100) and method integrate image preprocessing, deep metric feature extraction, prototype gallery construction, adaptive threshold optimization, and similarity matching into a unified workflow. An input layer (102) receives cattle muzzle images, which are processed by a preprocessing hub (104) for normalization, enhancement, and quality filtering. A deep feature extraction engine (106) uses a deep metric learning network to extract discriminative feature embeddings and generate L2-normalized query vectors. A similarity matching engine (110) computes cosine similarity between the query vector and a dynamic prototype gallery module (108). An adaptive threshold optimization engine (112) calculates dynamic similarity thresholds based on intra-class and inter-class prototype statistics, allowing an open-set decision engine (114) to distinguish enrolled cattle from unseen animals. An incremental enrollment engine (116) enables few-shot registration of new cattle identities without retraining the neural network.
Disclaimer: Curated by HT Syndication.
