Using a VTON model for real time production purpose
#34
by
aniket2025
- opened
I am searching a way, in which I can use the VTON model inside an app for real-time production purpose. I came to know that to do this, I have to pre-process my images with model image, cloth image, segmentation masks, pose estimation key points etc. However, the client wants if I can use any pre-trained model on my images and test the result on 10 images.
How can I do that without any UIs like ComfyUI or Stable Diffusion WebUI. I am not finding any concrete source of creating this model in my local system. I have tried to use Pose estimation + key points detection + Grounded SAM + Thin Plate Spline algorithm to achive the result. But, I am not successful in this yet.
Any advice/code/reference you can provide or suggest will help me a lot