TopoTTA: Topology-Enhanced Test-Time Adaptation for Tubular Structure Segmentation

2025-08-01

AI Summary

Tubular structure segmentation (TSS) suffers from incomplete segmentation and topological disconnections in cross-domain settings due to topological variations and local feature shifts. To address this, we propose TopoTTA—the first test-time adaptation (TTA) framework tailored for TSS. Its core contributions are: (1) Topological Meta-Difference Convolution (TopoMDC), which explicitly encodes tubular structural priors to enforce topological consistency; (2) Topological Hard-sample Generation (TopoHG), synthesizing pseudo-breakpoint regions to improve model robustness at topologically fragile locations; and (3) a prediction consistency optimization and pseudo-label alignment mechanism guided by clDice. TopoTTA operates without updating pretrained model parameters and is plug-and-play compatible with diverse CNN-based segmentation architectures. Evaluated across four cross-domain scenarios and ten datasets, it achieves an average clDice improvement of 31.81%, significantly mitigating topological degradation.

Technology Category

• Computer Vision: Segmentation
• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning
• Constraint Satisfaction and Optimization: Satisfiability Modulo Theories

Application Category

• Search and Retrieval-Augmented AI: Vertical and domain-specific search
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

Tubular structure segmentation (TSS) is important for various applications, such as hemodynamic analysis and route navigation. Despite significant progress in TSS, domain shifts remain a major challenge, leading to performance degradation in unseen target domains. Unlike other segmentation tasks, TSS is more sensitive to domain shifts, as changes in topological structures can compromise segmentation integrity, and variations in local features distinguishing foreground from background (e.g., texture and contrast) may further disrupt topological continuity. To address these challenges, we propose Topology-enhanced Test-Time Adaptation (TopoTTA), the first test-time adaptation framework designed specifically for TSS. TopoTTA consists of two stages: Stage 1 adapts models to cross-domain topological discrepancies using the proposed Topological Meta Difference Convolutions (TopoMDCs), which enhance topological representation without altering pre-trained parameters; Stage 2 improves topological continuity by a novel Topology Hard sample Generation (TopoHG) strategy and prediction alignment on hard samples with pseudo-labels in the generated pseudo-break regions. Extensive experiments across four scenarios and ten datasets demonstrate TopoTTA's effectiveness in handling topological distribution shifts, achieving an average improvement of 31.81% in clDice. TopoTTA also serves as a plug-and-play TTA solution for CNN-based TSS models.

Problem

• Address domain shifts in tubular structure segmentation
• Enhance topological representation without altering parameters
• Improve topological continuity with pseudo-label hard samples

Innovation

• Topological Meta Difference Convolutions enhance representation
• Topology Hard sample Generation improves continuity
• Plug-and-play TTA solution for CNN models