MUMBAI, India, July 30 -- Intellectual Property India has published a patent application (202641088898 A) filed by Indian Institute Of Technology Hyderabad on July 21, 2026, for System And Method For Automated Segmentation Of Tumor Regions From Multi-Modal Medical Imaging Data.
Inventors include Rahul Kumar; and Vinodini D.
The application for the patent was published on July 24, 2026, under issue no. 30/2026.
Abstract: A system and method for automated segmentation of tumor regions from multi-modal medical imaging (MI) data is disclosed. The method includes receiving, by an input processing unit (102), a plurality of MI modalities and generating a multi-modal input representation based on a plurality of image slices corresponding to the plurality of MI modalities. A feature extraction unit (104) processes the plurality of MI modalities to generate hierarchical feature representations at a plurality of spatial scales. A feature fusion unit (106) adaptively combines feature representations associated with the plurality of MI modalities to generate fused multi-modal feature representations. A spatial reasoning unit (108) generates a graph representation comprising region nodes and relational spatial connections, and performs graph-based aggregation to generate spatially encoded feature representations. A plurality of decoder units (110) generate segmentation feature maps, and an output generation unit (114) generates segmentation masks corresponding to one or more tumor regions. [Figure 1]
Disclaimer: Curated by HT Syndication.
