Yep, but how is that informative for a particular garment that you want to visualize? You can't infer, from a corpus of images of random clothes being worn, how

Yep, but how is that informative for a particular garment that you want to visualize? You can't infer, from a corpus of images of random clothes being worn, how that particular shape and fabric behaves.