Workflow Name: [Style Transfer Costume Change Face Swap] Portrait Photography
[Workflow Introduction]
This workflow achieves comprehensive image processing, including costume change, background replacement, face swap, and more. To start using this process, you need to upload three images: a body pose image, a clothing style image, and a facial image. Subsequently, the IPAdapter model is used for image style transfer, controlnet skeleton and depth models are used for pose control, and joy_caption is utilized to reverse-generate prompt text. These are connected to the sampler to generate the first image. Then, the controlnet depth model is used again, connected to the sampler for image generation, and the image is optimized. Next, facial cropping and face swap repainting are performed and overlaid onto the image, achieving the effect of a specified person wearing specific style clothing, enhancing the controllability of the entire process. You can also try this process to experience the fun and surprises brought by its powerful image processing capabilities!

[Usage Scenarios]
This process allows you to upload specified model, environment, and clothing images, then synthesize them to achieve the effect of a specified model wearing specific style clothing in a specified environment. Through detailed image processing and synthesis, it ensures the perfect integration of the model, clothing, and environment. This process helps you easily realize the visual effects you imagine, providing image outputs that meet your needs.

[Key Nodes]
IPAdapter, controlnet, SUPIR

[Model Version]
SDXL
Model Name: dreamshaperXL v21TuroDPMSDE.safctensors

[LoRA Model]
None

[ControlNet Application]
controlnet union sdxl loader
DW Pose Preprocessor
Start Time: 0
End Time: 0.5
Strength: 0.85

Depth Anything V2
Start Time: 0
End Time: 0.8
Strength: 0.85

[K Sampler]
CFG: 6.6
Sampling Method: euler_ancestral
Scheduler: normal
Denoising: 1