Introduction
This ComfyUI workflow is designed to achieve the effect of undressing a person. The workflow starts with an image of the person, which can be uploaded as a real photo or generated using an AI image generation model. Then a segmentation model is used to separate the person from the background and mark the clothing and body areas, generating masks for the clothing and body. Then flux and redraw nodes are used to remove the clothes, and finally prompt words are used to enhance the pose generation and the effect of removing the clothes.
Workflow Overview
How to Use This Workflow?
Step 1: Upload Pictures
1.Upload Pictures.
Step 2: Enter prompt words and adjust mask
1.Enter the prompt word, you can add a new prompt word at the end of the original prompt word.If the pose of the person in the generated image changes, you can add prompt words to further control the pose.
2.If the generated image has problems such as incorrect posture or incomplete characters, these problems may be caused by the mask covering too small an area or too large an area. You can adjust the 'expand' parameter. The larger the value, the larger the mask area, and the smaller the value, the smaller the mask area.
Step 3: Keep the accessories
1.In the parameter prompt, enter the name of the accessory you want to save. If there is no accessory, leave it blank.