Inpainting Specialist for Image Restoration
Budget: €12 – €18 EUR
We have a project that consists in swapping the person of some images, for a consistent LoRA character in HD
We have to find a workflow that is better to the one they have here: https://vmake.ai/fashion-model-studio that consists in segmenting the clothes, masking it, and generating by inpainting the model with certain controlnets in order to have consistency.
This technique uses OpenPose + IpAdapter + Segmentation in order to mantain the clothes intact and generate the virtual person outpainting around the clothes.
We need to use a LoRA of the virtual person to mantain the same person (same face, same hair lenght, same skin tone) all the time even if there are changes on the pose of the base image.
Can you generate a workflow with Automatic1111 or ComfyUI to create this type of images?
Here's an example of before / after: https://cdn.discordapp.com/attachments/1137484322245443685/1193300793755975780/image.png?ex=65ac36f7&is=6599c1f7&hm=830e4db44c9575245b95ee6ace1b37a9c2014e3a2a61015f98c76f5a735c99f7&
We have to find a workflow that is better to the one they have here: https://vmake.ai/fashion-model-studio that consists in segmenting the clothes, masking it, and generating by inpainting the model with certain controlnets in order to have consistency.
This technique uses OpenPose + IpAdapter + Segmentation in order to mantain the clothes intact and generate the virtual person outpainting around the clothes.
We need to use a LoRA of the virtual person to mantain the same person (same face, same hair lenght, same skin tone) all the time even if there are changes on the pose of the base image.
Can you generate a workflow with Automatic1111 or ComfyUI to create this type of images?
Here's an example of before / after: https://cdn.discordapp.com/attachments/1137484322245443685/1193300793755975780/image.png?ex=65ac36f7&is=6599c1f7&hm=830e4db44c9575245b95ee6ace1b37a9c2014e3a2a61015f98c76f5a735c99f7&