I tried to use segmentation models to extract out the human out of the pic; but I am not sure if this will scale; + segmentation has a max-size problem which make it hard to create a 1:1 image; to extract out the object! I am trying to create a segmentation image which I will use to subtract out the background;