cs.CVNov 15, 2026

Morphology-Decoupled Segmentation for Multiple Tunnel Defects Under Complex Environments

Authors: Mohan Ren, Hang Li, Yifei Liu, Shaowei Wang, James O'Kane, Xuanzhi Wang, Shiyu Zhao, Yumeng Zhang, 

Organizations: School of Civil Engineering, The University of Queensland, St. Lucia, Brisbane, 4072, Queensland, Australia · Department of Computer Science, University of Exeter, The Queen’s Drive, Exeter, EX4 4QJ, Devon, United Kingdom · School of Civil and Environmental Engineering, Nanyang Technological University, , Singapore, 639798, Singapore · School of Civil Engineering, Chang’an University, , Xi’an, 710064, Shaanxi, China · 

Abstract

Automated tunnel-lining inspection must segment different defect geometries: cracks are thin, elongated, and connectivity-sensitive, whereas leakage, lining falling off, and segment damage are spatially extensive and context-dependent. Existing approaches generally apply a common decoder across defect classes or introduce boundary, topology, and directional mechanisms as isolated enhancements. This paper introduces Morphology-Decoupled Segmentation (MDS), in which defect morphology becomes the organising principle for feature routing, decoding, and interaction. A supervised routing module partitions multiscale encoder features into complementary linear and areal streams. The linear stream employs Deformable Strip Attention with image-conditioned direction hypotheses to preserve crack continuity without assuming fixed horizontal or vertical orientations, while the areal stream combines Pyramid Pooling with feature-pyramid fusion to capture diffuse boundaries and regional context. The representations are coupled through gated cross-morphology attention, allowing complementary information exchange while retaining morphology-specific features. A resolution-adaptive tiling and probability-blending pipeline preserves fine defects across variable-resolution imagery. MDS achieves 81.17% mIoU on the internal validation set and improves foreground image-level recognition over a backbone-matched SegDINO model, primarily for areal defects. Controlled ablations demonstrate the complementary contributions of routing supervision and cross-stream interaction while revealing a trade-off between areal-class performance and crack specialisation. The submitted model achieved 67.55% mIoU on an organiser-controlled hidden test, and the team placed second. Selective dynamic INT8 quantisation reduced model storage by 2.83 times with negligible accuracy loss.

Figures & tables

Explore similar work

CardsList
  1. Training-Free Tunnel Defect Inspection and Engineering Interpretation via Visual Recalibration and Entity Reconstruction

    Apr 30, 2026Shipeng Liu, Liang Zhao, Dengfeng Chen +1Open-Vocabulary Object DetectionPromptable Image Segmentation

  2. Rethinking Efficient Crack Segmentation with Task-Aligned Structural-Directional Modeling

    May 29, 2026Shipeng Liu, Liang Zhao, Dengfeng Chen +1Efficient Semantic SegmentationCrack Segmentation

  3. MixerCSeg: An Efficient Mixer Architecture for Crack Segmentation via Decoupled Mamba Attention

    Mar 2, 2026Zilong Zhao, Zhengming Ding, Pei Niu +2Crack SegmentationHybrid CNN-Transformer Architectures