MUMBAI, India, June 26 -- Intellectual Property India has published a patent application (202641072413 A) filed by Sanskrithi School Of Engineering on June 11, 2026, for Real-Time Suspicious Activity Detection System For Video Surveillance.
Inventors include Dr. K. Vinod Kumar; Bhoomireddy Sreya; Akula Joshna; Bottu Varsha Vardhini; Kasireddy Nikhitha; and Gunda Mounika.
The application for the patent was published on June 19, 2026, under issue no. 25/2026.
Abstract: The present invention relates to a real-time suspicious activity detection system for video surveillance configured to automatically monitor live CCTV video streams and identify abnormal human behaviour using deep learning and computer vision techniques. The system comprises a video acquisition module, preprocessing module, YOLOv8-based object detection layer, Deep SORT object tracking module, behavioural analysis engine, face recognition layer, alert generation module, event logging system, heatmap visualization module, and Flask-based monitoring dashboard. The surveillance system continuously processes video frames captured from CCTV cameras, detects and tracks individuals, analyses behavioural patterns, and identifies suspicious activities including loitering, intrusion, abnormal movement, and violence. Facial recognition mechanisms are additionally integrated for identifying individuals involved in suspicious events. Upon detection of suspicious behaviour, the system automatically generates alerts, records evidence, and transmits notifications to authorized users. The invention further supports CSV-based event logging, heatmap generation, and web-based remote monitoring functionality. The disclosed invention provides improved surveillance efficiency, automated threat detection, reduced dependency on manual monitoring, enhanced operational transparency, and scalable deployment capability suitable for modern security and public-safety infrastructures.
Disclaimer: Curated by HT Syndication.
