MUMBAI, India, Jan. 2 -- Intellectual Property India has published a patent application (202541123490 A) filed by Anurag Engineering College, Suryapet, Telangana, on Dec. 8, 2025, for 'deep learning-powered facial recognition system for high-accuracy identification and security applications.'
Inventor(s) include Mrs. Jujjuru Vijayasree; and Mrs. G. Prashanthi.
The application for the patent was published on Jan. 2, under issue no. 01/2026.
According to the abstract released by the Intellectual Property India: "Modern facial recognition technologies have achieved exceptional accuracy through advancements in machine learning. At BARC's Computer Division, an indigenous deep learning-based facial recognition system has been created, capable of accurately identifying more than 20,000 individuals, even under real-world, unconstrained conditions. This article outlines the end-to-end development workflow of the system, demonstrating each stage of a full machine learning project. It covers essential considerations, including the selection of evaluation metrics, the choice and preparation of datasets, and the development of effective neural network architectures to maximize recognition performance. The discussion also extends to deployment strategies, emphasizing the use of edge computing to deliver fast, real-time inference without relying on centralized servers. Furthermore, the article highlights the increasing importance of data-centric AI practices, which prioritize enhancing data quality and uniformity. It also stresses the need for ongoing model monitoring and periodic updates to maintain dependable and high-performing recognition capabilities over long-term operational use."
Disclaimer: Curated by HT Syndication.
