I am trying to develop a DeepIllumination in directML.

DeepIllumination :
https://github.com/CreativeCodingLab/DeepIllumination?tab=readme-ov-file

directML :
https://github.com/microsoft/DirectML

However, the DeepIllumination model takes multiple images as input, and I need to stitch them into a single image.

So I want to stitch the rendered result images (color, normal, and depth map) into a single image and use it as input to the DeepIllumination model in real-time.

Do you know how to stitch these images into a single image in direct3D12?

Thanks.