tfm.vision.augment.wrap

Returns 'image' with an extra channel set to all 1s.