CLP: A Real-World Dataset of Contaminated Lens Protectors for Robust Semantic Segmentation

Sungyong Park1, Sooyoung Choi1, Hyunseo Koh1, Youngjae Choi1, Heewon Kim1,2
1 Soongsil University    2 Kairoba Inc.
CVPR 2026
Paper Abstract Dataset Experiments BibTeX

Abstract

The reliability of autonomous systems in real-world environments is mainly dependent on the robustness of their visual perception. Although recent studies have advanced the handling of visual degradations, physical contaminants that adhere to the camera lens, such as mud, water droplets, and condensation, remain largely underexplored. To this end, we introduce the CLP (Contaminated Lens Protector) dataset, a real-world benchmark designed to evaluate perception performance under realistic lens-protector contamination. The CLP dataset offers degraded images across multiple types of contamination and various lens-to-protector distances, along with dense semantic segmentation masks and aligned restoration targets.

Dataset

Overview Placeholder Add dataset overview text here later.
  • paired clean / degraded images
  • dense semantic labels
  • aligned restoration targets
Statistics Placeholder Add key numbers and dataset stats here later.
  • number of images
  • number of classes
  • distance settings
Release Placeholder Add download / access / benchmark release notes here later.
  • download link
  • benchmark protocol
  • leaderboard or submission URL

Experiments

Quantitative Results Placeholder Add benchmark tables, metrics, and ablations here later.
Qualitative Results Placeholder Add result figures, comparisons, and failure cases here later.

Conclusion

Conclusion Placeholder Add the final paper conclusion here later.
Limitation / Future Work Placeholder Add limitations and future directions here later.

Dataset Access

Download Placeholder Put dataset download or access-request link here later.
Submission Placeholder Put benchmark submission or leaderboard link here later.
Code Placeholder Put training / evaluation code link here later.

BibTeX

@inproceedings{park2026clp,
  title     = {CLP: A Real-World Dataset of Contaminated Lens Protectors for Robust Semantic Segmentation},
  author    = {Sungyong Park and Sooyoung Choi and Hyunseo Koh and Youngjae Choi and Heewon Kim},
  booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition},
  year      = {2026}
}
Copy BibTeX