ABSTRACT
A fundamental distinction between Gunnar Johansson’s research and much other contemporary research on vision involves the representation of the optical information for vision. In Gunnar’s research, the structure of visual stimulation is regarded as based on spatio-temporal relationships intrinsic to the optical patterns themselves. Gunnar’s experiments elegantly illustrate how space, motion, and shape should be regarded as interrelated aspects of visual stimulation.
One hypothesis suggested by Gunnar’s research on visual vector analysis and perception of biological motion is that information about the global structure and motion of motion configurations is provided by local optical properties. Indirect support for this hypothesis is provided by several recent experiments on the discriminability of globally coherent and incoherent motion, but efforts to identify specific local optical properties that carry such global information have not yet succeeded.
The general hypothesis that perceived spatial structure is governed by intrinsic properties of moving patterns is examined in (a) recent theoretical and experimental research on the perception of surface structure from motion, and (b) research on the “optic sphere theory.” Currently available evidence indicates that vision is very sensitive to information about surface shape provided by the local differential structure of moving images. Gunnar’s hypothesis that visual information for locating environmental objects and the observer within a 3D environmental framework is given by the structure of the “optic sphere” deserves continued experimental attention.
