Shape-Adapting Gated Experts: Dynamic Expert Routing for Colonoscopic Lesion Segmentation

2025-11-23

AI Summary

To address computational redundancy and limited model adaptability arising from high morphological heterogeneity of cells in whole-slide images (WSIs), this paper proposes SAGE, a dynamic expert routing framework. Methodologically, SAGE introduces a dual-path CNN-Transformer hybrid architecture, incorporates a Shape-Adaptive Hub (SA-Hub) to orchestrate multi-level shared and task-specific experts, and employs Top-K sparse activation for input-dependent dynamic inference. The resulting SAGE-UNet achieves state-of-the-art Dice scores of 95.57%, 95.16%, and 94.17% on the EBHI, DigestPath, and GlaS colonoscopy lesion segmentation benchmarks, respectively. Its design enables both superior accuracy and strong cross-domain generalization, demonstrating robustness across diverse histopathological domains and staining protocols.

Technology Category

• Machine Learning: Mixture of Experts (MoE)
• Computer Vision: Segmentation
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Graph Algorithms and Modeling for the Web: Algorithms and analysis for heterogeneous, signed, attributed, multi-relational, temporal, higher-order, and annotated Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

The substantial diversity in cell scale and form remains a primary challenge in computer-aided cancer detection on gigapixel Whole Slide Images (WSIs), attributable to cellular heterogeneity. Existing CNN-Transformer hybrids rely on static computation graphs with fixed routing, which consequently causes redundant computation and limits their adaptability to input variability. We propose Shape-Adapting Gated Experts (SAGE), an input-adaptive framework that enables dynamic expert routing in heterogeneous visual networks. SAGE reconfigures static backbones into dynamically routed expert architectures. SAGE's dual-path design features a backbone stream that preserves representation and selectively activates an expert path through hierarchical gating. This gating mechanism operates at multiple hierarchical levels, performing a two-level, hierarchical selection between shared and specialized experts to modulate model logits for Top-K activation. Our Shape-Adapting Hub (SA-Hub) harmonizes structural and semantic representations across the CNN and the Transformer module, effectively bridging diverse modules. Embodied as SAGE-UNet, our model achieves superior segmentation on three medical benchmarks: EBHI, DigestPath, and GlaS, yielding state-of-the-art Dice Scores of 95.57%, 95.16%, and 94.17%, respectively, and robustly generalizes across domains by adaptively balancing local refinement and global context. SAGE provides a scalable foundation for dynamic expert routing, enabling flexible visual reasoning.

Problem

• Dynamic expert routing addresses cellular heterogeneity in colonoscopic lesion segmentation
• Static CNN-Transformer hybrids cause redundant computation limiting input adaptability
• Shape-adapting framework balances local refinement and global context across domains

Innovation

• Dynamic expert routing for adaptive segmentation
• Hierarchical gating selects shared and specialized experts
• Shape-Adapting Hub bridges CNN and Transformer modules