双分支拉氏门控图聚合多模态脑膜瘤分割方法
By using a bi-branch Lagrange gating graph aggregation method, semantic alignment and spatial focusing of multimodal features are achieved, and edge information is explicitly restored. This solves the problems of feature misalignment and edge information loss in meningioma segmentation, and improves segmentation precision and anatomical continuity.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- SHANDONG UNIV OF SCI & TECH
- Filing Date
- 2026-04-28
- Publication Date
- 2026-07-17
AI Technical Summary
Existing technologies suffer from problems such as misalignment of multimodal heterogeneous features, loss of edge information due to downsampling, and imprecise and fragmented meningioma segmentation results due to lack of topological constraints.
A dual-branch Laplacian gated graph aggregation method is adopted. By combining a dual-branch multimodal input module and a hierarchical Swin Transformer encoder with a Laplacian gated graph aggregation module and an edge-aware feature modulation module, semantic alignment and spatial focusing of multimodal features are achieved, edge information is explicitly recovered, and global topological constraints are applied.
It improves the precision and anatomical continuity of meningioma segmentation, solves the problems of blurred boundaries and missed detection of small lesions, and ensures the anatomical integrity and computational efficiency of the segmentation results.
Smart Images
Figure CN122116088B_ABST