(preamble: I think this fits either here or SO with computer vision tag, I chose gamedev because I think you guys are probably really good with textures and meshes, but tell me if my decision was wrong.)
I want to accomplish the following: Given an image with an object, a 3D mesh very similar to that object, and rendering parameters that render that mesh to the exact location of this object. I want to "extract" the original texture from the image onto the mesh. (So in a later step, I could re-render the mesh from another viewpoint (of course some of the texture would be invisible/black)). My mesh has uv-coordinates for each triangle so I guess I could either "store the texture on the mesh" somehow, or directly backmap it to a 2D texture map.
So I guess what I want to do is kind of texture backmapping/remapping, kind of the inverse of what is done in game-dev when texturing objects. I was having a lot of trouble finding any useful information on Google about what I want to do, so I thought I'd ask here. Maybe I haven't found the right word for it yet. I think there's probably quite a lot of "stuff" involved because pixel in the original image won't exactly correspond to a location on the mesh.