Research map / Image generation / GANs
Towards Vivid and Diverse Image Colorization With Generative Color Prior
Yanze Wu, Xintao Wang, Yu Li, Honglun Zhang, Xun Zhao, Ying Shan
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 pretrained BigGAN generative color prior (ablation) | GCP-Colorization · ImageNet val | FID: 8.40 -> 3.62 | Improved |
| add pretrained BigGAN generative color prior (ablation) | GCP-Colorization · ImageNet val | Colorfulness: 31.21 -> 35.13 | Improved |
| add pretrained BigGAN generative color prior (ablation) | GCP-Colorization · ImageNet val | Delta Colorfulness: 7.14 -> 3.22 | Improved |
| use multi-resolution GAN features instead of inversion image guidance (ablation) | GCP-Colorization · ImageNet val | FID: 4.01 -> 3.62 | Improved |
| use multi-resolution GAN features instead of inversion image guidance (ablation) | GCP-Colorization · ImageNet val | Colorfulness: 26.12 -> 35.13 | Improved |
| use multi-resolution GAN features instead of inversion image guidance (ablation) | GCP-Colorization · ImageNet val | Delta Colorfulness: 12.23 -> 3.22 | Improved |
| add spatial alignment of GAN features (ablation) | GCP-Colorization · ImageNet val | FID: 4.59 -> 3.62 | Improved |
| add spatial alignment of GAN features (ablation) | GCP-Colorization · ImageNet val | Colorfulness: 31.94 -> 35.13 | Improved |
| add spatial alignment of GAN features (ablation) | GCP-Colorization · ImageNet val | Delta Colorfulness: 6.41 -> 3.22 | Improved |
| use last 7 discriminator layers or add pixel loss for BigGAN inversion instead of last 3 layers (ablation) | GCP-Colorization · BigGAN inversion (qualitative) | inversion quality | Hurt |
| add L2 norm penalty on latent code z in encoder training (ablation) | GCP-Colorization · BigGAN inversion (qualitative) | inversion quality | Improved |
| compare feed-forward GCP-Colorization against PSNR-oriented ranking | GCP-Colorization vs CIC · ImageNet val | PSNR: 20.86 -> 21.81 | No clear effect |
Closest papers by meaning
- BigColor: Colorization Using a Generative Color Prior for Natural Images (ECCV 2022 · similarity 0.91)
- PalGAN: Image Colorization with Palette Generative Adversarial Networks (ECCV 2022 · similarity 0.91)
- Learning Diverse Image Colorization (CVPR 2017 · similarity 0.90)
- DDColor: Towards Photo-Realistic Image Colorization via Dual Decoders (ICCV 2023 · similarity 0.88)
- Colorful Image Colorization (ECCV 2016 · similarity 0.88)
- HistoGAN: Controlling Colors of GAN-Generated and Real Images via Color Histograms (CVPR 2021 · similarity 0.88)
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.
