Face deformation workflow
About this model
Wan Move and Wan Vace two passes workflow
A workflow for rendering deformed faces in a two stage rendering pipeline with a intermediate pass of rendering
Inputs
Pass 1
- Input frame
- Height
- Width
- Frames (need to be 4n+1) (ie 33, 65, 81)
- prompt 1: simple prompt for text generation, note that it's not that useful.
Once you render this frame you need a second pass of a mask to rerender the region of the image that doesn't fit
Pass 2
- Pass 2 Mask : video of the first input with a zonning mask in medium grayscale (127, 127, 127) of the zone you want to rerender. No gaussian blur, sharp mask.
- Compare frame : just a frame you want to compare between the input and the generated mask on the workflow. If they don't matches, the result may be broken.
- prompt 2 : simple prompt for text generation, note that it's not that useful.
Workflow
- Load your image and input parametter
- change width, height, prompt and frames count
- disable 01_Pass 2
- execute the node next to the spline editor to start the editor with your reference image as background.
- click New Canvas to clean the spline.
- Right click Add new spline and ajust according to your reference image.
- Right click Add new single point to add an part of the image that will not move
- run the workflow
- look at the output in > deformation/temp
- Create a mask in medium grayscale (127, 127, 127) of the zone you want to rerender. No gaussian blur, sharp mask.
- Load the video in Pass 2 Mask and adjust the prompt
- disable 00_Pass 1 and enable 01_Pass 2
- run the workflow
Related Models
Similar AI models you may like
ON-THE-FLY 实时生成!Wan-AI 万相/ Wan2.1 Video Model (multi-specs) - CausVid&Comfy&Kijai - workflow included
【WAN2.1】IMG to VIDEO
ComfyUI Image Workflows
WAN 2.2 Workflow T2V-I2V-T2I (Kijai Wrapper)
Hunyuan 🌻 AllInOne
Wan2.1_14B_FusionX
Moody Simple Zimage Turbo/Distilled Workflow