Research map / Label-efficient and robust learning / Recognition and long-tail
CLCC: Contrastive Learning for Color Constancy
Yi-Chen Lo, Chia-Che Chang, Hsuan-Chao Chiu, Yu-Hao Huang, Chia-Ping Chen, Yu-Lin Chang, Kevin Jou
Check 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 CLCC contrastive learning with Full-Aug to SqueezeNet-FC4 | SqueezeNet-FC4 · NUS-8 | mean angular error: 2.23 -> 1.84 | Improved |
| Add CLCC contrastive learning with Full-Aug to SqueezeNet-FC4 | SqueezeNet-FC4 · NUS-8 | worst-25% angular error: 5.15 -> 4.20 | Improved |
| Add CLCC contrastive learning with Full-Aug to SqueezeNet-FC4 | SqueezeNet-FC4 · Gehler | mean angular error: 1.65 -> 1.44 | Mixed |
| Add CLCC contrastive learning with Full-Aug to SqueezeNet-FC4 | SqueezeNet-FC4 · Gehler | worst-25% angular error: 3.78 -> 3.48 | Improved |
| Use neutral-color WB-Aug instead of no CLCC augmentation (ablation) | SqueezeNet-FC4 · NUS-8 | mean angular error: 2.23 -> 1.93 | Improved |
| Use neutral-color WB-Aug instead of no CLCC augmentation (ablation) | SqueezeNet-FC4 · NUS-8 | worst-25% angular error: 5.15 -> 4.30 | Improved |
| Use full-color mapping (Full-Aug) instead of neutral-color mapping (WB-Aug) (ablation) | CLCC on SqueezeNet-FC4 · NUS-8 | mean angular error: 1.93 -> 1.84 | Improved |
| Use full-color mapping (Full-Aug) instead of neutral-color mapping (WB-Aug) (ablation) | CLCC on SqueezeNet-FC4 · NUS-8 | worst-25% angular error: 4.30 -> 4.20 | Improved |
| Add CLCC contrastive learning | SqueezeNet-FC4 · Gehler sparse illuminant cluster (16 data points) | worst-case error: 3.24 -> 2.31 | Improved |
Closest papers by meaning
- Cascading Convolutional Color Constancy (AAAI 2020 · similarity 0.88)
- Ranking-Based Color Constancy With Limited Training Samples (TPAMI 2023 · similarity 0.87)
- Multi-Domain Learning for Accurate and Few-Shot Color Constancy (CVPR 2020 · similarity 0.87)
- White-Balance First, Adjust Later: Cross-Camera Color Constancy via Vision-Language Evaluation (CVPR 2026 · similarity 0.86)
- GCC: Generative Color Constancy via Diffusing a Color Checker (CVPR 2025 · similarity 0.85)
- NightCC: Nighttime Color Constancy via Adaptive Channel Masking (CVPR 2024 · similarity 0.85)
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.
