

HD, details, simple, automatic
Principle:
1. Based on the FLUX FILL REDUX universal migration process, solves the issue of blurry details in reference images.
2. This workflow extracts and enlarges parts of the reference image to allow the FILL model to pay more attention to local details. Then, it crops out the areas of the main image that need to be redrawn, enlarges them for redrawing, with the goal of sampling and redrawing with more pixels.
3. The redrawn image is then scaled back to its original size and stitched together, thereby restoring more details of the reference image.
4. Face swapping, outfit changing, scene substitutions, and model replacements are all possible.
Operation Method:
5. Upload the main edited image, and in the input box, specify what you want to replace. For example, to replace clothes, write "clothes"; to replace a person, write "humen." Then, in the reference image prompt, specify the content you want to reference. For example, if referencing clothes, write "clothes"; if referencing a person, write "humen." Then click start.
6. You can set the number of output images. "1" means 1 image, "2" means 2 images, "4" means 4 images. The larger the number, the longer the processing time; generally, 1 image takes about 1 minute. The first run may take longer to load the model. Try multiple times to achieve the desired effect.
7. This workflow supports uploading a manually masked main edited image. If no mask is applied, the default is to use the cutout node.
8. If you have any questions, you can consult VX: FDS LXC
HD, details, simple, automatic
Principle:
1. Based on the FLUX FILL REDUX universal migration process, solves the issue of blurry details in reference images.
2. This workflow extracts and enlarges parts of the reference image to allow the FILL model to pay more attention to local details. Then, it crops out the areas of the main image that need to be redrawn, enlarges them for redrawing, with the goal of sampling and redrawing with more pixels.
3. The redrawn image is then scaled back to its original size and stitched together, thereby restoring more details of the reference image.
4. Face swapping, outfit changing, scene substitutions, and model replacements are all possible.
Operation Method:
5. Upload the main edited image, and in the input box, specify what you want to replace. For example, to replace clothes, write "clothes"; to replace a person, write "humen." Then, in the reference image prompt, specify the content you want to reference. For example, if referencing clothes, write "clothes"; if referencing a person, write "humen." Then click start.
6. You can set the number of output images. "1" means 1 image, "2" means 2 images, "4" means 4 images. The larger the number, the longer the processing time; generally, 1 image takes about 1 minute. The first run may take longer to load the model. Try multiple times to achieve the desired effect.
7. This workflow supports uploading a manually masked main edited image. If no mask is applied, the default is to use the cutout node.
8. If you have any questions, you can consult VX: FDS LXC