Abstract
It is generally believed that the main principle for the system for low level vision is so called bottom-up processing and it operates rather passively. It is not really true. The information processing for low-level vision is realized by interactions between the bottom-up and the top-down information flows. Especially, the sub-systems involved in size and depth perception functions only when proper assumptions or constraints are provided. These are good examples of active aspects of the low-level visual system. In regard of psychological realities, the kind of perception that involves more active processing, such as the perception of subjective contour with Kanizsa figure may have higher psychological reality or impact to us. This idea is yet a speculative hypothesis, but it is worth while to pursue.
