MUMBAI, India, April 17 -- Intellectual Property India has published a patent application (202641043580 A) filed by Vellore Institute Of Technology, Vellore, Tamil Nadu, on April 6, for 'a real-time vision-based indoor navigation and obstacle avoidance system for visually impaired users using deep learning and geometric scene analysis.'
Inventor(s) include Archana T; Siddharth Y; Raghul G; Jeshurun Nehemiah A; and Kanishka S J.
The application for the patent was published on April 17, under issue no. 16/2026.
According to the abstract released by the Intellectual Property India: "Real-Time Vision-Based Indoor Navigation and Obstacle Avoidance System Using Deep Learning and Geometric Scene Analysis. The present invention relates to a real-time vision-based indoor navigation system configured to assist visually impaired users by integrating deep learning-based object detection with geometric scene analysis. The system utilizes a monocular camera to capture continuous video frames of an indoor environment and processes the frames through an object detection module to identify obstacles and generate bounding box coordinates, object class labels, and confidence scores. A geometric scene analysis module detects structural elements including walls, doors, and staircases using image processing techniques. The system further estimates relative distance using a bounding box width-to-frame ratio and determines directional positioning based on centroid analysis. A centralized decision engine prioritizes obstacle detection and generates context-aware navigation instructions, which are converted into real-time audible output using Text-to-Speech technology. The system operates without reliance on external infrastructure, enabling cost-effective and portable deployment."
Disclaimer: Curated by HT Syndication.
