Bottom-up processing is a data-driven approach to perception where sensory input is processed starting from the smallest, most basic units and building up to a complete perception. It contrasts with top-down processing, which uses prior knowledge and expectations to interpret sensory information.