MUMBAI, India, Jan. 2 -- Intellectual Property India has published a patent application (202541122676 A) filed by S R University, Warangal, Telangana, on Dec. 5, 2025, for 'tumor segmentation using hierarchical multi-scale transformer on multi-modal mri data.'
Inventor(s) include Naga Maha Lakshmi K; and Durgesh Nandan.
The application for the patent was published on Jan. 2, under issue no. 01/2026.
According to the abstract released by the Intellectual Property India: "The present invention discloses a method and system for automatic segmentation of brain tumors from multi-modal magnetic resonance imaging (MRI) data. The invention employs a Hierarchical Multi-Scale Transformer Network (HiMAST-Net) comprising a Cross-Attentive Transformer Encoder (CATE) to capture intra- and inter-modality dependencies across T1, T2, FLAIR, and T1ce scans. An Adaptive Windowed Context Aggregation (AWCA) module dynamically adjusts receptive fields based on local complexity, while a Sparsity-Aware Transformer (SAT) reduces redundant computations by suppressing background regions. A Modality-Attentive Feature Pyramid Decoder (MAFPD) generates an initial segmentation mask using attention-prioritized skip connections. Segmentation accuracy is enhanced by an Uncertainty-Guided Segmentation Refinement (UGSR) process and a Refined Edge Boundary Smoothing (REBS) mechanism for improved boundary delineation. The invention achieves precise identification of tumor sub-regions, including enhancing tumor, tumor core, and edema, with improved computational efficiency, thereby supporting clinical diagnosis, treatment planning, and neuro-oncology research."
Disclaimer: Curated by HT Syndication.
