Creating sets of images with slight variations (happy, sad, angry, tired, etc.)

Creating a set of multiple images that differ only by specified features would be extremely valuable and efficient for myself and others working in the field of autism and early learning.  When teaching a learner to discriminate and identify a specific item or adjective, it can be a barrier if the stimuli differ too much.

For example, 5 nearly identical images of a raccoon that differ only in the facial expression of the raccoon can better help the learner to identify adjectives (happy, sad, angry, tired, scared, surprised, etc.).