My stuff

  • My Workflows

  • Liked Workflows

  • Following Workflows

Go to OpenArt main site
Upload workflow

Automatically draw couple photos with customized backgrounds, faces and clothing

5.0

0 reviews
12
5.0K
1.1K
1
Description

Have your clothes, characters and background composition ready.

(fix with new ipadapter)

Start setting up a couple photo today.

SEGS (Ordered) is used to automatically generate Colormask, so there is no need to draw the mask manually.


Follow the Steps:

1.Upload your image.

At least six pictures are required, including two portraits (marked by Face, red and green blocks), one background (marked by Background, blue blocks), and two clothes (marked by Clothes, yellow and light blue blocks). One action (marked by Pose)

2.Base imagesize

Change the size of your base image based on the long side. The size of the base image is based on your pose image.

3.Setting up Colormask

Set the Take_Start and Take_Count of SEGS Filter(Ordered)node, to determine the position of your face Colormask, the combination is (0,1) or (1,2).

For example, Take_Start is 0 and Take_Count is 1

Or Take_Start is 1 and Take_Conut is 2

A switch function will be added here in the future.

4.Setting up Conditions

Enter your character conditions in the light blue and yellow areas. The background conditions are set in the Loader. The character conditions should preferably include descriptions of clothes and hairstyles. You can use WD14 Tagger to summarize the text to make the description more accurate.

5.Setting up Ipadapter ClipVision

Adjust your image position for ClipVision so that the ipAdapter has a more precise reference

Then Start Queue


==

If the generated image does not meet your expectations, please confirm that your Colormask and Pose are correct

SDXL should also work normally,Please modify all the Ipadapter node and Controlnet node.

The regional ColorMask cannot currently be used with Insightface, so FaceID is not used in the workflow.

Discussion

(No comments yet)

Loading...

Author

67
129.9K
2.0K
558.0K

No reviews yet

  • - latest (2 years ago)

  • - v20240203-042543

Primitive Nodes (4)

Note (1)

workflow/controlnet (1)

workflow/紅色區域條件 (1)

workflow/綠色區域條件 (1)

Custom Nodes (44)

ComfyUI

  • - LoadImage (6)

  • - InvertMask (1)

  • - PreviewImage (5)

  • - ConditioningCombine (2)

  • - DifferentialDiffusion (1)

  • - InpaintModelConditioning (1)

  • - CLIPTextEncode (1)

  • - MaskToImage (2)

ComfyUI Impact Pack

  • - SegsToCombinedMask (5)

  • - ImpactSEGSOrderedFilter (2)

  • - UltralyticsDetectorProvider (2)

  • - BboxDetectorSEGS (1)

  • - SegmDetectorSEGS (1)

  • - LayerMask: PersonMaskUltra (1)

  • - IPAdapter (5)

  • - IPAdapterUnifiedLoader (1)

  • - IPAdapterUnifiedLoaderFaceID (1)

  • - Efficient Loader (1)

  • - KSampler (Efficient) (2)

  • - GetImageSize (1)

  • - JWImageResizeByLongerSide (1)

Checkpoints (0)

LoRAs (0)