(This template is used for Workflow Contest)

What this workflow does

👉 In this workflow we try and explore one concept of making T shirt mockups with some cool Input images and using the IP adaptor to convert same into final images.  Input images can be any AI art generated or your own collection of images that you want to overlay over the model .

How to use this workflow

👉Model can be taken from any opensource free resources of which plenty are available.

In this workflow we try and merge two masks one from "clipseg" and another from Mask inpainting so that the combined mask acts as a place holder for image generation.

The idea is sometimes the area to be masked may be different from the semantic segment by clipseg and also the area may not be properly fixed by automatic segmentation.

Ensure the finished mask which is shown int he preview matches the needs and if not adjust same by additional masks and also adjust the required params in clipseg including blur factors to match and get a good quality mask.

The quality of the end result totally depends on the style of the mask and form factor.

Tips about this workflow

👉 I have kept the blur values low here as this particular set of examples need a share outline rather than a blurred outline.

Normally the image addition will take care of the alpha needs and if required additional alpha channels can be added.

Also note this workflow can be extended not only for T shirts but for other objects also where you need to create a mockup run and of which many are existing.

These mockups can be used as examples in Print on demand websites or ETSY platforms etc to showcase the product display.

Resolution can be enhanced by using suitable up scalers in the pipelines.

More info can be had once you are satisfied with the outcome and can be enhanced to make it custom templates for your needs. If any information on same is required do ping us on @jags111

🎥 Video demo link (optional)

👉 [Please add here]