A technique that divides images into small square regions (patches) for processing, rather than analyzing individual pixels.
Quality of vision, audio, and image understanding (distinct from modality support)