MUMBAI, India, Nov. 2 -- Intellectual Property India has published a patent application (202641125670 A) filed by Dr. K. Manoharan on October 22, 2026, for An Uncertainty-Aware, Resource-Adaptive Edge Ai System For Early Brain Tumor Detection, Multi-Region Segmentation And Classification Using Multisequence Mri.
Inventors include Dr. Logaprakash M; Dr. Suresh Kumar N; Mr. Krishna Yadav Kattagoni; Dr Vaibhav-Shah; Dr Manoharan K; Madhav Kumar Mishra; Dr. Mohmmad Ashaq Sofi; Mr G Narendran; Mr. K. B. Lingkash; Mr. Hari Prakash R; Mr. Sommath Teli; and Dr Rajendiran M.
The application for the patent was published on October 30, 2026, under issue no. 44/2026.
Abstract: The invention relates to an uncertainty-aware, resource-adaptive edge artificial intelligence system for analyzing multisequence brain magnetic resonance imaging data. The system comprises an MRI input interface, preprocessing module, sequence availability manager, feature extraction and fusion network, segmentation and classification modules, uncertainty estimation module, resource monitoring module, adaptive inference controller, and refinement module. The system generates prediction outputs and associated uncertainty indicators, obtains a resource state of an edge computing device, and selects an inference configuration according to the resource state and predefined inference requirements. When an output fails a configured acceptance criterion, the controller selectively initiates additional processing subject to resource constraints and a stopping condition. The system distinguishes predictive uncertainty from input-quality limitations and may generate spatial uncertainty maps, segmentation overlays, classification outputs, and processing records. The architecture enables resource-aware local MRI analysis and identifies outputs requiring additional assessment.
Disclaimer: Curated by HT Syndication.
