Workflow Name: Video Mask Area Redrawing
【Workflow Introduction】
Upload a video at the "load video" node, and upload an image as a style reference at the "load image" node. This workflow will extract the characters from the video, then pass them to the sampler for redrawing through ControlNet's open pose, depth, ip2p, ip adapter, and other nodes, while keeping the background unchanged. Click run and give it a try!

【Use Cases】
For example, if you upload a dance video and an artwork as a style reference, after processing through this workflow, the masked area of the dancer in the video will be redrawn according to your reference style image, creating a unique and stunning visual effect. Come and experience this magical workflow to bring a new artistic charm to your video.

【Key Nodes】
controlnet, ip adapter, AnimateDiff Loader

【Model Version】
SD1.5
Model Name: majicmixRealistic v7.safetensors

【LoRA Model】
Model Name: sd15 lora beta AnimateLCM.safetensors

【ControlNet Application】
Depth Preprocessor
Start Time: 0
End Time: 1
Strength: 0.41

IP2P Preprocessor
Start Time: 0
End Time: 1
Strength: 0.67

openpose Preprocessor
Start Time: 0
End Time: 1
Strength: 0.39


【K Sampler】
CFG: 8
Sampling Method: lcm
Scheduler: sgm uniform
Denoising: 1