Synapse
⌘+K
Synapse
PulseExploreClubsResearchersJournals
Instagram
HomeClubsExplore
May 24, 2026Applied SciencesOpen Access

Crack Segmentation Model for Low-Quality Crack Images Based on Feature Integration and Triple Attention

View Full Paper
Ask AI
Bookmark
Share

Authors

YXYonghua XieYWYue Wang

Discussion

Loading...

Member takes

Overview

Randomized trial demonstrates improved crack detection in low-quality images, suggesting better practical applications.

Key Points

  • The study aims to improve road crack detection in low-quality images using a novel segmentation model.
  • Proposed a crack segmentation model based on feature integration and triple attention mechanism.
  • Used DeepLabv3+ as the backbone network with additional three-dimensional interactive attention and multi-group dilation modules.
  • Evaluated model performance on Crack500 and GAPS384 datasets.
  • Achieved superior segmentation performance compared to existing methods, especially for small and weak cracks.
  • Successfully reduced missed detections in low-quality images, enhancing overall accuracy.
  • Maintained practical inference efficiency without excessive model size.

Cite This Study

Xie et al. (2026) studied this question.

synapsesocial.com/papers/6a1296b248a0ea1665673aafhttps://doi.org/10.3390/app16115185
View Full Paper
Ask AI
Bookmark
Share

Also Consider

Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context:

  1. 1Training Surface Crack Segmentation Networks with Group-wise Normalization Attention and Padding-Shifting-Cutting Convolution2024 · 6 citations
  2. 2Robust Semantic Segmentation for Automatic Crack Detection Within Pavement Images Using Multi-Mixing of Global Context and Local Image Features2024 · 49 citations
  3. 3A lightweight feature attention fusion network for pavement crack segmentation2024 · 17 citations
  4. 4LiteCrackSeg: A lightweight hybrid CNN-transformer for efficient crack segmentation.2026 · 2 citations
  5. 5Research on road crack segmentation based on deep convolution and transformer with multi-branch feature fusion.2024 · 9 citations