My stuff

  • My Workflows

  • Liked Workflows

  • Following Workflows

Go to OpenArt main site
Upload workflow

LivePortrait Omni(LivePortrait 全能版)

5.0

0 reviews
151
20.7K
5.0K
10
Description

这套工作流总共提供三种方式,主推IMG2VID以及VID2VID,如果你想要IMG2VID,请在切换节点选择1,如果你想要VID2VID,请选择2。另外你想进行实时换脸,可以选择这两个节点中任意一个将其变为Driving Video。

提供3种Face Detectors工具,Insightface:严格用于非商业用途,其他两个未明确限制商业用途。在效果上,face-alignment和Insightface相较于MediaPipe更出色。

顺便说一下,如果主体是人,Model选择human;如果主体是动物,Model选择animal。

重点!重点!重点!:

使用前请看Youtube以及Bilibili视频,你会得到一切你意料之外的收获!!!这值得你帮我点赞( •̀ω •́ )y,这对我将来继续开源清晰有用的工作流以及极富营养的教程有很大支持!!!


Youtube:Jason Stone

https://youtu.be/bMgIwvKLo_0

Bilbili:JasonSt0ne

https://www.bilibili.com/video/BV1oNYxe3EgV


         LivePortrait官网:https://liveportrait.github.io/

LivePortrait(Kijai)插件:https://github.com/kijai/ComfyUI-LivePortraitKJ

LivePortrait(Kijai)模型下载:https://huggingface.co/Kijai/LivePortrait_safetensors/tree/main

LivePortrait在线Demo:https://huggingface.co/spaces/KwaiVGI/LivePortrait

最推荐的人脸识别插件:https://github.com/1adrianb/face-alignment

素材寻找网站:

https://www.istockphoto.com/ (收费质量高)

https://www.pexels.com/videos/(免费居多)

裁剪视频比例网站:

https://online-video-cutter.com/


This workflow provides a total of three methods. IMG2VID and VID2VID are mainly recommended. If you want IMG2VID, please select 1 in the switching node. If you want VID2VID, please select 2. In addition, if you want to perform real-time face swapping, you can select either of these two nodes to turn it into a Driving Video.

3 Face Detectors tools are provided, Insightface: is strictly for non-commercial use, and the other two do not explicitly restrict commercial use. In terms of effect, face-alignment and Insightface are better than MediaPipe.

By the way, if the subject is a human, the Model selects human; if the subject is an animal, the Model selects animal.

The point! The point! The point! :

Please watch the YouTube and Bilibili videos before using this tool. You will gain insights beyond your expectations! Your thumbs-up is greatly appreciated ( •̀ω •́ )y. It significantly supports my efforts to continue open-sourcing clear and practical workflows, as well as creating deeply informative tutorials in the future!


Youtube:Jason Stone

https://youtu.be/bMgIwvKLo_0

Bilbili:JasonSt0ne

https://www.bilibili.com/video/BV1oNYxe3EgV

LivePortrait official website: https://liveportrait.github.io/

LivePortrait (Kijai) plug-in: https://github.com/kijai/ComfyUI-LivePortraitKJ

LivePortrait (Kijai) model download: https://huggingface.co/Kijai/LivePortrait_safetensors/tree/main

LivePortrait online demo: https://huggingface.co/spaces/KwaiVGI/LivePortrait

The most recommended face recognition plug-in: https://github.com/1adrianb/face-alignment

Material search website:

https://www.istockphoto.com/ (high quality charge)

https://www.pexels.com/videos/ (mostly free)

Cropped video ratio website:

https://online-video-cutter.com/

Discussion

(No comments yet)

Loading...

Author

9
14.5K
307
57.3K
    animal.png (554.4 kB)
    human.jpg (52.1 kB)

No reviews yet

  • - latest (a year ago)

Primitive Nodes (24)

DF_Image_scale_to_side (2)

DownloadAndLoadLivePortraitModels (1)

Fast Groups Bypasser (rgthree) (1)

LivePortraitComposite (1)

LivePortraitCropper (1)

LivePortraitLoadCropper (1)

LivePortraitLoadFaceAlignmentCropper (1)

LivePortraitLoadMediaPipeCropper (1)

LivePortraitProcess (1)

Note (7)

PrimitiveNode (1)

Screencap_mss (1)

WebcamCaptureCV2 (1)

easy getNode (2)

easy setNode (2)

Custom Nodes (17)

ComfyUI

  • - PreviewImage (4)

  • - LoadImage (1)

ComfyUI Essentials

  • - MaskPreview+ (2)

  • - ImpactSwitch (1)

  • - VHS_LoadVideo (2)

  • - VHS_VideoCombine (1)

  • - GetImageSizeAndCount (4)

  • - CreateShapeMask (1)

  • - GrowMaskWithBlur (1)

Checkpoints (0)

LoRAs (0)