MUMBAI, India, May 1 -- Intellectual Property India has published a patent application (202621026435 A) filed by Prof. Nilam Honmane; Dishant Nitin Parjane; Dhawal Narendra Parkhe; Harshal Satish Patil; and Purushottam Vijaykumar Patil, Pune, Maharashtra, on March 6, for 'autonomous multi-modal crowd surveillance system with audio-visual fusion and predictive risk forecasting.'
Inventor(s) include Prof. Nilam Honmane; Dishant Nitin Parjane; Dhawal Narendra Parkhe; Harshal Satish Patil; and Purushottam Vijaykumar Patil.
The application for the patent was published on May 1, under issue no. 18/2026.
According to the abstract released by the Intellectual Property India: "This invention presents a multi-mode surveillance system, the Proposed system AI, which is an autonomous system, used to improve safety of people in the crowded urban setting. The system combines the computer vision based on deep learning (YOLOv8) and real-time audio spectral analysis (FFT) to form an Echo-Visual fusion architecture. This architecture identifies panic audio signatures (2-4kHz) and intelligently enhances visual sensitivity to detect threats in blind spots or obstructed locations. Besides, the system has a Future-Cast engine, which uses a Social Force Model to model the crowd dynamics and estimate the risk of congestion at least two minutes in advance. The system will have automated forensic reporting features, which will produce immutable PDF reports with embedded telemetry as soon as an incident is detected. A rate of experimental validation shows that there is an improvement of the detection rate of 82 mod 94 percent in occlusion conditions and also real-time processing in edge hardware."
Disclaimer: Curated by HT Syndication.
