ComfyUI-FBCNN

ComfyUI-FBCNN
★ 37

图像修复JPEG去伪影去块处理ComfyUI节点
在ComfyUI中调用FBCNN对JPEG图片进行去伪影与去块修复,恢复细节并提升画质,易于集成与运行。
💡 对压缩严重的JPEG图片进行去伪影与画质恢复。
🍴 6 Forks💻 Python🔄 2025-02-24
📦
网盘下载
复制链接后前往夸克网盘下载
https://pan.quark.cn/s/6862a2001521
📄 README

ComfyUI-FBCNN

ComfyUI implementation of FBCNN

What is it?

This is a JPEG artifact/compression removal tool, which can work automatically or with the help of a user.

Installation

Download or git clone this repository inside ComfyUI/custom_nodes/ directory. The only required dependencies are Pytorch, numpy, and requests, which should be already installed if you have ComfyUI.

How to use it?

In ComfyUI, you can find this node under image/upscaling category.

The model requires an estimate of the compression level, which is a number between 0 and 100 (the same you need to provide when compressing a JPEG image). You can either set auto_detect to enabled and use built-in estimator, or set it to disabled and provide the value yourself in compression_level.

⚠️ If you leave auto_detect as enabled, the value of compression_level will be ignored.

Credits

  • FBCNN
  • ComfyUI