A PSYCHOPHYSICAL INVESTIGATION OF A DYNAMIC DEPTH PERCEPTION THRESHOLD IN THE HUMAN VISUAL SYSTEM USING ROTATING THREE-DIMENSIONAL LISSAJOUS PATTERNS

Abstract

A tridimensional optical display is described which is capable of producing real images with no optics between the observer and the image. This display, which is called the Visually Integrated Volumetric Image (VIVID), was developed as a laboratory system designed to be used in vision research. Three dimensional Lissajous patterns were produced, whose rotational position and axial depth could be precisely controlled by a general purpose digital minicomputer. By reversing the direction of rotation of the pattern in a known manner, and recording a subject's ability to perceive the true rotational direction for various amounts of axial image depth, the transition between the kinetic depth effect and true depth perception could be investigated. This transition forms the basis for defining a depth perception threshold for this type of a dynamic three dimensional image. Two experiments were performed using this approach. In the first experiment, ten undergraduate and graduate students (both male and female) observed the pattern during 24 one-minute trials. Eight different axial depth values were used, with three trials at each depth. The patterns were made to reverse rotation direction in a pseudo-random manner which was different for each trial presented to a given subject. The subject indicated the perceived direction of rotation by controlling a two position rotary switch, which was monitored by the computer. The total time during which the subject correctly tracked the rotation was integrated by the computer. The ratio of this correct tracking time to the total trial time was taken as a measure of the subject's ability to perceive the axial depth of the image. Plots of this ratio as a function of image depth are presented and discussed. The results averaged over all subjects produce a remarkably smooth curve, yielding a depth perception threshold of 2 mm for the subject population under the given experimental conditions. The characteristics of the stimulus were: (a) 2 cm high by 2 cm wide pattern; (b) Lissajous pattern frequency ratio of 6:1; (c) Axial distance from observer of 55 cm; (d) Wavelength of 533 nm (green); (e) Horizontal rotation axis; (f) Rotation period of 7 seconds; (g) Neutral background field illumination over a 40° come angle; (h) Image line width of 0.5 mm. The second experiment was a pilot experiment designed to investigate the effect of wavelength variations on this form of depth perception. A single subject was presented with a total of 192 one-minute trials at four wavelengths spanning the visible region (i.e., 6 one-minute trials for each of eight depth values at each wavelength). No significant wavelength dependence was observed in this pilot experiment.