Research map / Detection and segmentation / Open-set and few-shot detection
CoGenSAM: Codebook-Interactive Generative Labeling for Adapting SAM to Crack Segmentation
Zhuangzhuang Chen, Nuo Chen, Dachong Li, Zhiliang Lin, Xingyu Feng, Yifan Zhang, Jie Chen, Jianqiang Li
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add the restoration model to edge-detection labels (ablation) | CoGenSAM · DeepCrack | mIoU: 49.22 -> 61.73 | Improved |
| add the restoration model to edge-detection labels (ablation) | CoGenSAM · Bridge2025 | mIoU: 32.78 -> 48.25 | Improved |
| add the pre-trained crack-aware codebook (ablation) | CoGenSAM · DeepCrack | mIoU: 57.33 -> 61.73 | Improved |
| add the pre-trained crack-aware codebook (ablation) | CoGenSAM · Bridge2025 | mIoU: 43.98 -> 48.25 | Improved |
| add contrastive-interaction loss Lc (ablation) | CoGenSAM · DeepCrack | mIoU: 58.01 -> 61.73 | Improved |
| add contrastive-interaction loss Lc (ablation) | CoGenSAM · Bridge2025 | mIoU: 44.52 -> 48.25 | Improved |
| add contrastive-interaction loss Lc (ablation) | CoGenSAM · DeepCrack | F1: 71.41 -> 74.35 | Improved |
| swap Sobel for Roberts, Prewitt or Canny edge detection (ablation) | CoGenSAM · DeepCrack | mIoU: 61.57 -> 61.34 / 61.00 / 60.41 | No clear effect |
| use generated labels instead of manual annotation (annotation-free vs CrackSAM) | SAM · DeepCrack | mIoU: 65.62 -> 61.57 | Hurt |
| use generated labels instead of manual annotation (annotation-free vs CrackSAM) | SAM · Crack500 | mIoU: 54.22 -> 46.33 | Hurt |
| use CoGenSAM generated labels instead of FreeCOS annotation-free method | SAM · Bridge2025 | mIoU: 41.05 -> 48.12 | Improved |
| use CoGenSAM instead of Up-CrackNet | SAM · Crack500 | mIoU: 21.27 -> 46.33 | Improved |
Closest papers by meaning
- Mind Marginal Non-Crack Regions: Clustering-Inspired Representation Learning for Crack Segmentation (CVPR 2024 · similarity 0.87)
- SCSegamba: Lightweight Structure-Aware Vision Mamba for Crack Segmentation in Structures (CVPR 2025 · similarity 0.85)
- Uncertainty-aware Fine-tuning of Segmentation Foundation Models (NeurIPS 2024 · similarity 0.84)
- Cracks in the Foundation: A Civil Infrastructure Dataset to Challenge Vision Foundation Models (NeurIPS 2026 · similarity 0.83)
- Scribbles for All: Benchmarking Scribble Supervised Segmentation Across Datasets (NeurIPS 2024 · similarity 0.82)
- Towards Open-Vocabulary Industrial Defect Understanding with a Large-Scale Multimodal Dataset (CVPR 2026 · similarity 0.82)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
