Research map / Image generation / Image restoration
CCMNet: Leveraging Calibrated Color Correction Matrices for Cross-Camera Color Constancy
Dongyoung Kim, Mahmoud Afifi, Dongyun Kim, Michael S. Brown, Seon Joo Kim
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 |
|---|---|---|---|
| condition on camera fingerprint embedding from calibrated CCMs instead of extra test-camera images (C5) | CCMNet · Gehler-Shi | mean angular error: 2.36 -> 2.23 | Improved |
| use CCMNet instead of C5 (m=9) | NUS-8 · NUS-8 | mean angular error: 2.54 -> 2.32 | Improved |
| use CCMNet instead of C5 on Cube+ | CCMNet · Cube+ | mean angular error: 1.87 -> 1.68 | Improved |
| remove need for additional test camera images; reduce model size | CCMNet vs C5 · Gehler-Shi | model size (MB): 2.09 -> 1.05 | Improved |
| add camera fingerprint embedding (CFE) to SIIE backbone (ablation) | SIIE · Cube+ | mean angular error: 3.39 -> 2.60 | Improved |
| add CFE to SIIE backbone on Gehler-Shi (ablation) | SIIE · Gehler-Shi | mean angular error: 3.67 -> 3.62 | No clear effect |
| add imaginary camera augmentation plus CFE to SIIE (ablation) | SIIE · Cube+ | mean angular error: 3.39 -> 1.91 | Improved |
| add CFE encoder to backbone (ablation) | CCMNet · Gehler-Shi | mean angular error: 2.53 -> 2.23 | Improved |
| add CFE encoder without augmentation (ablation) | CCMNet · Gehler-Shi | mean angular error: 2.79 -> 2.74 | No clear effect |
| use interpolated imaginary camera augmentation (0<=alpha<=1) instead of alpha=1 camera mapping (ablation) | CCMNet · Cube+ | mean angular error: 1.86 -> 1.68 | Improved |
| add augmentation (alpha=1) without CFE on Gehler-Shi (ablation) | backbone f · Gehler-Shi | mean angular error: 2.79 -> 2.87 | Hurt |
Later papers that compared with this method
1 later accepted paper reports CCMNet as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- Multi-Block-Attention-based Color Constancy (ECCV 2026 · better on 4 of 4 shared results)
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- When Color Constancy Goes Wrong: Correcting Improperly White-Balanced Images (CVPR 2019 · similarity 0.90)
- White-Balance First, Adjust Later: Cross-Camera Color Constancy via Vision-Language Evaluation (CVPR 2026 · similarity 0.90)
- Cross-Camera Convolutional Color Constancy (ICCV 2021 · similarity 0.89)
- Attentive Illumination Decomposition Model for Multi-Illuminant White Balancing (CVPR 2024 · similarity 0.89)
- Single Pixel Spectral Color Constancy (IJCV 2023 · similarity 0.88)
- GCC: Generative Color Constancy via Diffusing a Color Checker (CVPR 2025 · similarity 0.87)
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.
