The question is whether this actually "creates a 3d model based on the picture", or if it "finds an existing model that looks similar to the picture and texture map it".