Absolute and relative effects of similarity and distance on grouping

Four experiments are reported, the aim of which was to explore the achromatic-colour and distance relations that determine the probability of achromatic surfaces grouping perceptually. The first two experiments were performed to test Wertheimer's conjecture that similarity and dissimilarity of achromatic colours jointly determine grouping. The results indicate that only similarity of achromatic colours determines grouping. Data from the other two experiments show that grouping is determined by absolute, relative, and Gillam's distances. These findings agree with previous literature showing that different algorithms or mechanisms determine grouping by similarity and grouping by distance. Additionally, these findings show that the single factor of grouping by distance also depends on multiple algorithms or mechanisms.     

The influence of texture on judgments of slant and relative distance in a picture with suggested depth

Six surfaces from natural environments with different visual textures were photographed at angles of 60, 65, and 70 deg from perpendicular. Measurements were taken of 24 Ss’ judgments of the inferred angles of slant and inferred midpoints of the six textured surfaces represented in the photographs which were viewed in the frontoparallel plane. Judgments of both slant and relative distance within the photographs were influenced by represented angle of slant and by variations in surface texture.     

Perceived relative distance on the ground affected by the selection of depth information

Our visual space does not appear to change when we scan or shift attention between locations. This appearance of stability implies that the depth information selection process is not crucial for constructing visual space. But we present evidence to the contrary. We focused on space perception in the intermediate distance, which depends on the integration of depth information on the ground. We propose a selection hypothesis that states that the integration process is influenced by where the depth information is selected. Specifically, the integration process inaccurately represents the ground when one samples depth information only from the far ground surface, instead of sequentially from the near to the far ground. To test this, observers matched the depth/length of a sagittal bar (test) to the width of a laterally oriented bar (reference) in three conditions in a full-cue environment that compelled the visual system to sample from different parts of the ground. These conditions had the lateral reference bar placed (1) adjacent to the test bar, (2) at the far ground, and (3) at the near ground. We found that the sagittal bar was perceived as shorter in conditions (1) and (2) than in Condition 3. This finding supports the selection hypothesis, since only Condition 3 led to more accurate ground surface integration/representation and less error in relative distance/depth perception. Also, we found that performances in all three conditions were similar in the dark, which has no depth information on the ground, indicating that the results cannot be attributed to asymmetric visual scanning but, rather, to differential information selection.     

Binocular depth discrimination depends on orientation

Using both the method of adjustment and forced-choice techniques, it was found that binocular depth thresholds depend on the orientation of the test targets, which in these experiments consisted of long, thin rods. Stereoacuity is greatest with vertical rods and decreases progressively as the angle of orientation approaches horizontal. The relationship between stereoacuity and orientation is governed by the orientation of the images of the targets on the retina, and is predicted fairly well by the sine of the angle of orientation, suggesting an equivalence between angular rotation and reductions in rod length.     

Effect of disparity and viewing distance on perceived depth

It is shown that veridical depth perception presupposes the processing of both the magnitude of retinal disparity and observation distance according to a square-law function specified by the underlying geometrical stimulus relations. In the present study, after testing its existence, this constancy of depth perception was investigated by measuring perceived depth as a function of retinal disparity and observation distance. In addition, the relative effectiveness of convergence and accommodation as possible indicators of distance was examined through a conflicting-cues paradigm. It was shown that in the perception of depth the visual system computes distance by taking into account the convergence parameter only, rather than that of accommodation or of both.     

Background surface and horizon effects in the perception of relative size and distance

The projected height of an object in a scene relative to a ground surface influences its perceived size and distance, but the effect of height should change when the object is moved above the horizon. In four experiments, observers judged relative size or relative distance for pairs of objects varying in height with respect to the horizon. Higher objects equal in projected size were judged larger below the horizon, but the relative size effect was reversed either when one object was on the horizon and one was above the horizon or when both objects were above the horizon. With the real horizon not explicitly present in the display, relative size judgements were affected both by the boundary of the visible surface and the vanishing point implied by the converging lines. For relative distance judgements, the higher object was judged more distant regardless of the height of the objects relative to the perceptual horizon, resulting in a reversal of the relation between size and distance judgements for objects above the horizon.     

Velocity gradients and relative depth perception

The effectiveness of velocity gradients in providing relative depth information was assessed using random dot patterns translating horizontally. The gradients simulated two planes meeting at a horizontal line at the center, and subjects judged whether the center was the nearest or farthest part of the display. Accuracy increased with maximum dot speed, exceeding 90% in conditions combining the highest speed (10.4^o/sec) and longer of two display durations (10 sec) with unrestricted fixation. Separate experiments examined a rotational component perceived in the motion of the planes and the latency in reporting a rigid organization of the displays. Possible reasons for the chance accuracy found by Farber and McConkie (1979) and alternative explanations of the effect of maximum dot speed on accuracy are discussed. A model is presented that accounts for the effects of dot speed and display duration on the accuracy of relative depth judgments.     

The effects of stereoscopic depth on completion

Stereoscopic depth has a critical effect on completion of partially occluded figures. However, it has not strictly been distinguished whether the effect is direct or indirect through alteration of contour segmentation or parsing. Here, I report that stereoscopic depth does not influence completion of partially occluded figures when parsing is unambiguous from motion cues. This is consistent with the present proposal that stereoscopic depth does not have a unique role in completion and that it is one of the cues to contour segmentation or parsing, which in turn influences completion and surface representation, like motion, shape, or transparency.     

The influence of visual texture density gradients on relative distance judgements

A hypothesis derived from J. J. Gibson's psychophysical theory of space perception was tested. Subjects made monocular relative distance judgements by moving a marker to the apparent physical mid-point between two other fixed markers which were placed on a surface along the subjects' line of sight. Judgements were significantly influenced by the texture density gradients of stimulation derived from the surface over which they were made.     

Visual discrimination of local surface depth and orientation

Many theoretical analyses of 3-dimensional form perception assume that visible surfaces in the environment are perceptually represented in terms of local mappings of metric depth and/or orientation. Although this approach is often taken for granted in the study of human vision, there have been relatively few attempts to demonstrate its psychological validity empirically. In an effort to shed new light on this issue, our research has been designed to investigate the accuracy with which observers can discriminate metric depth and orientation intervals on smoothly curved surfaces. Observers were presented with visual images of surfaces defined by shading and/or texture, on which two pairs of points were designated with small dots. In Experiment 1, their task was to identify which pair of points had a greater difference in depth; in Experiment 2 they were required to judge which pair had a greater difference in orientation. The Weber fractions obtained for these tasks were 10 to 100 times greater than those that have been reported for other types of sensory discrimination, indicating that the perception of metric structure from these displays is surprisingly coarse grained.     

The perception of size and distance under monocular observation

A relative-perceived-size hypothesis is proposed to account for the perception of size and distance under monocular observation in reduced-cue settings. This hypothesis is based on two assumptions. In primary processing, perceived size is determined by both proximal stimulation on the retina and distance information from primary cues such as oculomotor cues. In secondary processing, the relation of two primary perceived sizes determines another relation of secondary perceived distances, so that an object of smaller primary perceived size is judged to be further away. An experiment was designed to test this hypothesis, especially the assumption of secondary processing, by making ratio judgments of perceived size and perceived distance for two successively presented targets. The Standard square was presented at a constant distance and varied in visual angle; the variable square was presented with a constant visual angle in distance. The results showed that an inverse relation between size and distance estimates held regardless of whether the visual angles of the targets were the same or different.     

The effects of scopolamine on a two-cue discrimination

Rats were trained with a tone, light or a tone--light combination as the discriminative stimulus. These groups were tested after doses of scopolamine and it was found that groups trained with a single cue were more sensitive to the drug than double-cue groups, although their pre-drug responding was similar. A similar pattern was found among individuals in the double-cue groups in which there was a significant correlation between dependency on a single cue, as shown in transfer tests, and drug sensitivity. These results were interpreted in terms of scopolamine-induced changes in stimulus sensitivity produced by a modification of the neural mechanisms controlling attention.     

The effects of duration and luminance on binocular depth mixture

Two stimuli in the same binocular direction, one in front and the other an equal disparity behind a fixation point, are perceived at one depth. This depth is between that corresponding to the two stimulus disparities and varies continuously from one stimulus to the other as a function of the ratio of their luminances. When either duration or absolute luminance is increased, perceived depth changes toward the midpoint of the two disparities.     

Perception of relative depth interval: systematic biases in perceived depth

Given an estimate of the binocular disparity between a pair of points and an estimate of the viewing distance, or knowledge of eye position, it should be possible to obtain an estimate of their depth separation. Here we show that, when points are arranged in different vertical geometric configurations across two intervals, many observers find this task difficult. Those who can do the task tend to perceive the depth interval in one configuration as very different from depth in the other configuration. We explore two plausible explanations for this effect. The first is the tilt of the empirical vertical horopter: Points perceived along an apparently vertical line correspond to a physical line of points tilted backwards in space. Second, the eyes can rotate in response to a particular stimulus. Without compensation for this rotation, biases in depth perception would result. We measured cyclovergence indirectly, using a standard psychophysical task, while observers viewed our depth configuration. Biases predicted from error due either to cyclovergence or to the tilted vertical horopter were not consistent with the depth configuration results. Our data suggest that, even for the simplest scenes, we do not have ready access to metric depth from binocular disparity.     

Perception of relative depth interval: Systematic biases in perceived depth

Given an estimate of the binocular disparity between a pair of points and an estimate of the viewing distance, or knowledge of eye position, it should be possible to obtain an estimate of their depth separation. Here we show that, when points are arranged in different vertical geometric configurations across two intervals, many observers find this task difficult. Those who can do the task tend to perceive the depth interval in one configuration as very different from depth in the other configuration. We explore two plausible explanations for this effect. The first is the tilt of the empirical vertical horopter: Points perceived along an apparently vertical line correspond to a physical line of points tilted backwards in space. Second, the eyes can rotate in response to a particular stimulus. Without compensation for this rotation, biases in depth perception would result. We measured cyclovergence indirectly, using a standard psychophysical task, while observers viewed our depth configuration. Biases predicted from error due either to cyclovergence or to the tilted vertical horopter were not consistent with the depth configuration results. Our data suggest that, even for the simplest scenes, we do not have ready access to metric depth from binocular disparity.