Abstract
A new orthodoxy in vision research has emerged from new insights into the structure of the visible environment and how this structure makes vision possible. These primary insights have radical implications for other aspects of visual theory. Machine vision research has used them to develop working models which span the traditional fields of visual sensory processes, spatial perception, form perception and recognition. Gibsonians have stressed other lines of development, notably the creation of a language for perceptual theory which avoids confusions embedded in older terminologies. They have often argued that contemporary machine vision systems differ substantially from human vision. However, this dispute does not affect the basic validity of either tradition: they can be understood as different and largely complementary ways of converging on a shared but distant goal, a theory incorporating models which are both full and faithful.

This publication has 21 references indexed in Scilit: