27 lines
833 B
Markdown
27 lines
833 B
Markdown
# ComfyUI-VideoColorGrading
|
|
|
|
ComfyUI implementation of [Video Color Grading via Look-Up Table Generation](https://github.com/seunghyuns98/VideoColorGrading) (ICCV 2025).
|
|
|
|
Generates a 3D color LUT from a reference image and source video frames using a two-stage diffusion process, then applies it for consistent color grading.
|
|
|
|
|
|
|
|
|
|
## Nodes
|
|
|
|
- **Load VCG Model** - Loads the combined model checkpoint (CLIP ViT-B/32, VAE, ReferenceNet, L-Diffuser)
|
|
- **Generate Color LUT (VCG)** - Generates a 16^3 3D LUT from reference image + source frames
|
|
- **Apply 3D LUT (VCG)** - Applies the generated LUT to images
|
|
|
|
## Setup
|
|
|
|
Model:
|
|
|
|
https://huggingface.co/Kijai/VCG_comfy/tree/main/checkpoints
|
|
|
|
|
|
## Links
|
|
|
|
- Paper: https://arxiv.org/abs/2508.00548
|
|
- Original code: https://github.com/seunghyuns98/VideoColorGrading
|