MUMBAI, India, July 30 -- Intellectual Property India has published a patent application (202641087133 A) filed by New Horizon College Of Engineering on July 16, 2026, for Adaptive Multi-Modal Edge Ai Framework For Self-Learning Industrial Defect Detection And Quality Assurance.
Inventors include Bhavya R; Kavitha K N; and Divyanshi Chhabra.
The application for the patent was published on July 24, 2026, under issue no. 30/2026.
Abstract: The present invention discloses an adaptive multi-modal edge AI framework for industrial defect detection and quality assurance. The framework comprises an image acquisition module configured to capture product images, a dynamic illumination compensation engine configured to generate region-specific correction maps, a convolutional feature extraction engine configured to derive visual feature vectors, and a graph-based defect relationship analyzer configured to evaluate spatial relationships among detected defect regions. A hybrid anomaly scoring engine computes anomaly scores using both visual feature deviation metrics and graph-topology deviation metrics, while a reinforcement-learning threshold optimizer continuously adapts inspection thresholds based on production feedback and validation outcomes. An industrial actuation interface communicates inspection results to manufacturing control systems and initiates corrective actions, and a production feedback repository stores operational data for continuous learning and optimization. The invention enables real-time detection of known and previously unseen defects, improves inspection accuracy under varying production conditions, reduces false-positive outcomes, and enhances overall industrial quality assurance efficiency.
Disclaimer: Curated by HT Syndication.
