MUMBAI, India, Aug. 10 -- Intellectual Property India has published a patent application (202641091428 A) filed by Pragati Engineering College on July 28, 2026, for A System And Method For Real-Time Computer Vision-Based Object Detection And Scene Understanding.
Inventors include Mr. L. Praveen Kumar; Mr. B. Sri Datta Subrahmanyam; Pallaprolu Satya Swaroop; Madhireddy Mani Deepika; and Kola Deva Siva Ganesh Bhargav.
The application for the patent was published on July 31, 2026, under issue no. 31/2026.
Abstract: ABSTRACT A System and Method for Real-Time Computer Vision-Based Object Detection and Scene Understanding The present disclosure provides a system and method for real time computer vision based object detection and scene understanding using adaptive multi-modal semantic reasoning, dynamic confidence calibration and context aware resource orchestration . The system captures visual data from one or more imaging sensors, pre-processes the captured data, harmonizes multi-scale visual features, extracts object characteristics, and dynamically adapts processing according to environmental conditions. The system also captures semantic relationships between detected objects, ensures object persistence over time, constructs a semantic scene graph, calibrates detection confidence, provides estimates of prediction uncertainty, eliminates false detections, retains contextual information through cross-frame semantic memory, and forecasts scene dynamics. An adaptive resource orchestration controller and a hardware-aware inference optimizer dynamically allocate computational resources to maintain real-time performance efficiency. The detected objects are verified by a multi-stage semantic verification engine and then an explainable decision generation module produces interpretable outputs. The disclosed invention improves detection accuracy, contextual awareness, computational efficiency, robustness in difficult operating conditions and reliability for intelligent vision applications.
Disclaimer: Curated by HT Syndication.
