Research map / Image generation / Image restoration
Colorful Image Colorization
Richard Zhang, Phillip Isola, Alexei A. Efros
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 class rebalancing to the classification loss (ablation) | Ours (class) · ImageNet test (10k images) | AuC rebalanced (%): 65.1 -> 67.3 | Improved |
| add class rebalancing to the classification loss (ablation) | Ours (class) · ImageNet test (10k images) | AMT labeled real (%): 25.2 -> 32.3 | Improved |
| add class rebalancing to the classification loss (ablation) | Ours (class) · ImageNet test (10k images) | AuC non-rebalanced (%): 91.6 -> 89.5 | Hurt |
| use classification loss instead of L2 regression, both trained from scratch (ablation) | Ours (L2) · ImageNet test (10k images) | AuC rebalanced (%): 64.4 -> 65.1 | Improved |
| use classification loss with rebalancing instead of L2 regression (ablation) | Ours (L2) · ImageNet test (10k images) | AMT labeled real (%): 21.2 -> 32.3 | Improved |
| fine-tune L2 network from the classification network instead of from scratch (ablation) | Ours (L2) · ImageNet test (10k images) | AuC rebalanced (%): 64.4 -> 66.2 | Improved |
| recolorize grayscale images with the method before VGG classification | VGG classifier · ImageNet test (10k images) | VGG top-1 class accuracy (%): 52.7 -> 56.0 | Improved |
| fine-tune VGG on grayscale inputs instead of recolorizing | VGG classifier · ImageNet test (10k images) | VGG top-1 class accuracy (%): 56.0 -> 63.5 | Improved |
| add color to the input of the VGG classifier (ground truth vs grayscale) | VGG classifier · ImageNet test (10k images) | VGG top-1 class accuracy (%): 52.7 -> 68.3 | Improved |
| our full method versus Larsson et al. on AMT realism | Ours (full) · ImageNet test (10k images) | AMT labeled real (%): 27.2 -> 32.3 | No clear effect |
| train on SUN-subset comparison: our ImageNet model versus Deshpande et al. | Ours (full) · SUN subset (Deshpande 240 images) | AMT labeled real (%): 9.8 -> 17.2 | Improved |
| use color input (Lab) in conv1 instead of grayscale for PASCAL segmentation | Ours · PASCAL VOC 2012 segmentation | mIoU (%): 35.0 -> 35.6 | Improved |
Closest papers by meaning
- Learning Representations for Automatic Colorization (ECCV 2016 · similarity 0.93)
- Colorization Transformer (ICLR 2021 · similarity 0.92)
- DDColor: Towards Photo-Realistic Image Colorization via Dual Decoders (ICCV 2023 · similarity 0.92)
- Pixelated Semantic Colorization (IJCV 2019 · similarity 0.92)
- Instance-Aware Image Colorization (CVPR 2020 · similarity 0.91)
- Automatic Controllable Colorization via Imagination (CVPR 2024 · similarity 0.89)
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.
