batchImg-rembg

Rembg(Remove background) of image sequence for ComfyUI </br></br>

<img src = 'https://github.com/Mamaaaamooooo/batchImg-rembg-ComfyUI-nodes/assets/135937372/d3e05963-b047-4900-aa58-10f1e1b0980c' width="400" height="400"></img> <img src = 'https://github.com/Mamaaaamooooo/batchImg-rembg-ComfyUI-nodes/assets/135937372/bef5f8b4-3976-4c59-848f-7e77df6bd5a3' width="400" height="400"></img>

Installation

1. Clone to your custom_nodes folder in ComfyUI:

git clone https://github.com/Mamaaaamooooo/batchImg-rembg-ComfyUI-nodes.git

2. Install rembg[gpu] (recommended) or rembg, depending on GPU support, to your ComfyUI virtual environment. E.g.:

pip install rembg[gpu]
pip install tqdm

batchImg-rembg workflows will often make use of these helpful node packs:

---

</br>

• ComfyUI-AnimateDiff-Evolved for img2vid.

• ComfyUI-VideoHelperSuite for loading videos, combining images into videos.

• comfyui_controlnet_aux for preprocessing original images to depth, lineart, openpose images.

  </br>

Workflow

for example,

<img src= 'https://github.com/Mamaaaamooooo/batchImg-rembg-ComfyUI-nodes/assets/135937372/a516f83b-f149-45be-9dba-31c35c719f3b'>

</br> </br> </br> <img src = 'https://github.com/Mamaaaamooooo/batchImg-rembg-ComfyUI-nodes/assets/135937372/17966afa-0b8a-4774-95d0-2c57b3846694'>

Optional Models(Choose according to your work, downloaded automatically)

---

All models are downloaded and saved in the user home folder in the .u2net directory.

The available models are:

<details open> <summary>Rembg Model Name </summary>

| Name | Description | Link | |--------|-------------------------------------------------------------------------------------|------------------------------------| | u2net(default) | A pre-trained model for general use cases. | download, source | | u2netp | A lightweight version of u2net model. | download, source | | u2net_human_seg | A pre-trained model for human segmentation. | download, source | | u2net_cloth_seg | A pre-trained model for Cloths Parsing from human portrait. Here clothes are parsed into 3 category: Upper body, Lower body and Full body. | download, source | | silueta | Same as u2net but the size is reduced to 43Mb. | download, source | | isnet-general-use | A new pre-trained model for general use cases. |download, source | | isnet-anime | A high-accuracy segmentation for anime character. | download, source | | sam(not recommended, not easy to use) | A pre-trained model for any use cases. | download encoder, download decoder, source |

</details> </br> </br>

Example

<table class="center"> <tr style="line-height: 0"> <td width=34% style="border: none; text-align: center">Original Image</td> <td width=33% style="border: none; text-align: center">LineArt before Rembg</td> <td width=33% style="border: none; text-align: center">LineArt after Rembg</td> </tr> <tr> <td width=34% style="border: none"><img src="https://github.com/Mamaaaamooooo/batchImg-rembg-ComfyUI-nodes/assets/135937372/ef065506-e844-4b92-a282-d20198267f8e" style="width:100%"></td> <td width=33% style="border: none"><img src="https://github.com/Mamaaaamooooo/batchImg-rembg-ComfyUI-nodes/assets/135937372/e4440bec-d6dd-4726-8200-c38facbd1130" style="width:100%"></td> <td width=33% style="border: none"><img src="https://github.com/Mamaaaamooooo/batchImg-rembg-ComfyUI-nodes/assets/135937372/0d828694-1440-464b-9c68-7f646b73886f" style="width:100%"></td> </tr> </table> </br>

Acknowledgements

Thanks to rembg-comfyui-node for insight.</br> Thanks to rembg-comfyui-node-better for modifying repo.