MUMBAI, India, Sept. 14 -- Intellectual Property India has published a patent application (202641106612 A) filed by Kkr & Ksr Institute Of Technology And Sciences on September 04, 2026, for A Real-Time Video-Based Surveillance Detection System Using Deep Learning For Intelligent Security Monitoring.
Inventors include Mr. Prof. Repudi Ramesh; Mrs. G. Srujana; Nalluri Sai Deepika; Pasupuleti Dunde Neelima; Pendem Saahithi; Pavuluri Jahnavi; Pasam Naga Parameswari; and T. Rajiya.
The application for the patent was published on September 11, 2026, under issue no. 37/2026.
Abstract: ABSTRACT [0013] A real-time video-based surveillance detection system using deep learning is disclosed for overcoming the limitations of traditional CCTV monitoring that depends on continuous human observation. Live video streams from surveillance cameras are processed frame by frame after preprocessing operations that include resizing, normalization, and noise reduction. A You Only Look Once (YOLO) object-detection model identifies individuals and relevant objects within each frame and produces bounding boxes with confidence scores. The DeepSORT multi-object tracking algorithm then assigns consistent identities to detected persons and constructs continuous movement trajectories across successive frames. Behavior-analysis techniques examine these trajectories to recognize suspicious patterns such as loitering in restricted zones, abnormal movement speeds or directions, and crowd formation. When a computed suspicion score exceeds a predefined threshold, the system automatically generates alerts that notify security personnel. The integrated pipeline improves surveillance efficiency, reduces operator fatigue and missed incidents, and provides a scalable solution suitable for airports, shopping areas, transportation hubs, and smart-city deployments.
Disclaimer: Curated by HT Syndication.
