MUMBAI, India, Sept. 22 -- Intellectual Property India has published a patent application (202641109522 A) filed by Mahendra College Of Engineering Autonomous on September 11, 2026, for Ai-Enabled Multimodal Thermal And Computer-Vision Surveillance System For Concealed-Object And Academic-Security Event Detection.
Inventors include Dr. M. Karthikeyan; Mrs. K. Priyadevi; Arul. C; Dhakshan A; Dhiyanesh N; Hari Pradhap S; Nandhaguru S; Rathik E; Ragul R; and Poovarasan A.
The application for the patent was published on September 18, 2026, under issue no. 38/2026.
Abstract: [059] The present invention discloses an AI-enabled multimodal thermal and computer-vision surveillance system for concealed-object and academic-security event detection. The system includes thermal and visible-spectrum sensors, an edge acquisition and synchronization module, a cross-modal registration and feature extraction module, a modality-reliability estimator, an adaptive AI fusion and risk-inference engine, a temporal verification module, an academic-security policy module, a privacy-aware event memory and an alert/evidence controller. Thermal and visible observations corresponding to a common physical region are registered and evaluated together with motion and contextual features. Respective modality-reliability values are determined from thermal contrast, illumination, occlusion, sensor health and registration confidence and are used to condition multimodal fusion, thereby limiting the influence of degraded sensor data. Candidate concealed-object or academic-security events are verified over time and correlated with zone-specific and time-specific institutional rules. Event evidence is selectively retained according to a risk tier, reducing continuous storage and communication while providing auditable event metadata and bounded evidence capture. Accompanied Drawing [FIGS. 1-2].
Disclaimer: Curated by HT Syndication.
