OpenArt
OpenArt
Workflows

    My stuff

  • My Workflows

  • Liked Workflows

  • Following Workflows

Go to OpenArt main site
Upload workflow

Vace Perfect Video Face change/enhance

5.0

0 reviews
11
3.0K
764
2
Description

Ever had that perfect video... but the face... is mush?  

Ya we have all been there, now you can fix it.

This workflow is set up as a tutorial flow.  Strait left to right.  Its long, but its the best ... and i do mean BEST video face change or enhancer that i have ever seen for videos.  The only downside is it cant fully keep lip syncing.

This uses wan 2.1 vace to cut out and enhance/change faces.

This changes both face and hair. Using either sam 1, or sam 2 masking, with simple bboxes or segmentation.

This allows full prompting of the face, allowing facial expressions to be added.

2 step process, one for hair and face, and one for just inner face.  Allowing HD detailed faces.

Sizes can be as small as 320x320 for cutouts up to 1024x1024 depending how large the face is in the video and your vram.


Leave a like if you like this.. i dont think you will find this disappoint you if you learn how to use it.


Discussion

(No comments yet)

Loading...

Author

That guy

That guy

36
62.5K
735
271.4K

No reviews yet

  • - latest (4 months ago)

Primitive Nodes (220)

Any Switch (rgthree) (5)

DepthAnythingV2Preprocessor (1)

DownloadAndLoadSAM2Model (2)

Fast Groups Bypasser (rgthree) (34)

Fast Groups Muter (rgthree) (8)

GetNode (60)

Image Batch To Sublist (mtb) (5)

ImageResizeKJv2 (9)

Lora Loader Stack (rgthree) (1)

MaskFix+ (6)

MaskPreview (2)

ModelSamplingSD3 (2)

Note (15)

PointsEditor (2)

PrimitiveNode (24)

PrimitiveStringMultiline (4)

ReActorUnload (2)

Reroute (6)

Sam2Segmentation (2)

SetNode (22)

TrimVideoLatent (2)

UnetLoaderGGUF (1)

UnetLoaderGGUFDisTorchMultiGPU (1)

WanVaceToVideo (2)

WanVideoVACEStartToEndFrame (2)

Custom Nodes (157)

Comfyroll Studio

  • - CR Model Input Switch (2)

  • - CR Image Input Switch (2)

ComfyUI

  • - ImageFromBatch (7)

  • - ImageScaleBy (1)

  • - VAELoader (1)

  • - LoraLoaderModelOnly (3)

  • - UNETLoader (1)

  • - UpscaleModelLoader (1)

  • - ImageBlend (1)

  • - InvertMask (4)

  • - EmptyImage (2)

  • - JoinImageWithAlpha (4)

  • - CLIPLoader (1)

  • - MaskToImage (5)

  • - PreviewImage (10)

  • - CLIPTextEncode (4)

  • - KSampler (2)

  • - FeatherMask (2)

  • - VAEDecodeTiled (2)

  • - LoadImage (2)

ComfyUI Essentials

  • - GetImageSize+ (1)

ComfyUI Impact Pack

  • - ImageListToImageBatch (1)

  • - SegsToCombinedMask (5)

  • - MaskListToMaskBatch (4)

  • - ImpactSimpleDetectorSEGS (5)

  • - UltralyticsDetectorProvider (5)

  • - ImpactGaussianBlurMask (2)

ComfyUI Nodes for Inference.Core

  • - DWPreprocessor (1)

ComfyUI-VideoHelperSuite

  • - VHS_DuplicateMasks (2)

  • - VHS_VideoInfoLoaded (1)

  • - VHS_VideoCombine (10)

  • - VHS_LoadVideoPath (1)

KJNodes for ComfyUI

  • - ImageUpscaleWithModelBatched (1)

  • - GetImageSizeAndCount (6)

  • - VRAM_Debug (7)

  • - ImageAndMaskPreview (6)

Masquerade Nodes

  • - Cut By Mask (10)

  • - Change Channel Count (16)

  • - Image To Mask (4)

  • - Paste By Mask (4)

  • - Mask To Region (4)

ReActor Node for ComfyUI

  • - ReActorRestoreFace (2)

WAS Node Suite

  • - Masks Combine Batch (2)

Checkpoints (0)

LoRAs (3)

wan\control\Wan21_AccVid_T2V_14B_lora_rank32_fp16.safetensors

wan\control\Wan21_CausVid_14B_T2V_lora_rank32_v1_5_no_first_block.safetensors

wan\control\Wan21_T2V_14B_MoviiGen_lora_rank32_fp16.safetensors