Stereopsis with chromatic signals from the blue-sensitive mechanism

The influence of stereopsis on visual saliency in a proto-object based model of selective attention

Some animals including humans use stereoscopic vision which reconstructs spatial information about the environment from the disparity between images captured by eyes in two separate adjacent locations. Like other sensory information, such stereoscopic information is expected to influence attentional selection. We develop a biologically plausible model of binocular vision to study its effect on bottom-up visual attention, i.e., visual saliency. In our model, the scene is organized in terms of proto-objects on which attention acts, rather than on unbound sets of elementary features. We show that taking into account the stereoscopic information improves the performance of the model in the prediction of human eye movements with statistically significant differences.

Effect of chromatic contrast on stereoacuity measurement with computer-aided three-dimensional technology

Background

Various measurement tools are utilized to detect the stereopsis threshold in the clinic, but seldom of these involves chromatic information. Incorporating colorful elements into computer-aided, three-dimensional (3D) evaluation systems could help the tests appear more vivid and attractive. The aim of this study was to clarify the effect of different chromatic pair stereo targets on the stereoacuity result.

Methods

A total of 17 subjects with visual acuity in each eye of at least 0 logarithmic minimum angle of resolution (logMAR) and a stereoacuity of at least 32 second of arc (arcsec) were recruited. A 3D laptop with liquid crystal shutter glasses was used for evaluating stereoacuity. Thirteen pages were set including seven maximum color contrast pages and six isoluminant color contrast pages.

Results

In maximum color contrast pair, no significant difference was found among all seven experiments [six test groups and one reference group, one-way analysis of variance (ANOVA) test, F=0.995, P=0.493]. There was also no significant difference among the isoluminant color contrast pairs (six test groups, one-way ANOVA test, F=0.873, P=0.644). Paired t-test was used for comparing the data between the same hue series in the maximum color contrast pair vs. the isoluminant color contrast pair, and significant differences were found in all six pairs (P<0.001).

Conclusions

Adding chromatic factors to the stereo test is practical to evaluate stereopsis.

Colour helps to solve the binocular matching problem

The spatial differences between the two retinal images, called binocular disparities, can be used to recover the three-dimensional (3D) aspects of a scene. The computation of disparity depends upon the correct identification of corresponding features in the two images. Understanding what image features are used by the brain to solve this binocular matching problem is an important issue in research on stereoscopic vision. The role of colour in binocular vision is controversial and it has been argued that colour is ineffective in achieving binocular vision. In the current experiment subjects were required to indicate the amount of perceived depth. The stimulus consisted of an array of fronto-parallel bars uniformly distributed in a constant sized volume. We studied the perceived depth in those 3D stimuli by manipulating both colour (monochrome, trichrome) and luminance (congruent, incongruent). Our results demonstrate that the amount of perceived depth was influenced by colour, indicating that the visual system uses colour to achieve binocular matching. Physiological data have revealed cortical cells in macaque V2 that are tuned both to binocular disparity and to colour. We suggest that one of the functional roles of these cells may be to help solve the binocular matching problem.

On the independence of chromatic and achromatic stereopsis mechanisms

The extent to which the processing of stereoscopic depth information can take place separately in colour-contrast-sensitive and luminance-contrast-sensitive mechanisms has been investigated. Contrast thresholds for stereoscopic depth identification (front/back) were measured using 0.5 c/deg Gabor patches. The stimuli possessed different amounts of colour and luminance contrast ranging from isoluminance (red/green) to isochrominance (yellow/black) through intermediate values. Two models for combining chromatic and achromatic stereopsis information were tested. The first (single-pathway) model assumed colour and luminance contrast summation within a single luminance-contrast-sensitive mechanism before stereoscopic judgement. The second (dual-pathway) model assumed probability summation between independent chromatic and achromatic stereopsis mechanisms. The latter model provided the better fit to the data. In providing evidence in favour of an independent chromatic stereopsis mechanism, it was shown that luminance artifacts were unlikely to be the cause of maintained stereopsis at isoluminance. The possible neural substrates of chromatic stereopsis are discussed.

Influence of the luminance and opponent chromatic channels on stereopsis with random-dot stereograms

The present work examines the relationship between random-dot stereograms (via the disparity range parameter) and color-vision mechanisms (via the luminance channel and red-green and tritan directions at isoluminance). The results clearly indicate that the variations in the stereograms along red-green confusion lines contribute to stereopsis. Stereoscopic perception depends on spatial information for stereograms generated with variations along tritan confusion lines. For observers who perceive stereopsis via tritan directions, the results show a gradation in the disparity range, with the disparity range for stereograms generated by luminance variations being greater than for stereograms generated in red-green directions; the latter range is, in turn, greater than for stereograms generated along tritan directions.

Binocular rivalry with isoluminant stimuli visible only via short-wavelength-sensitive cones

To test whether the binocular contour rivalry mechanism is tritanopic, we presented isoluminant, rival stimuli visible only via the short-wavelength-sensitive (S) cones. We stimulated only the S cones with violet gratings superimposed on a bright yellow field that adapted the responses of the middle- and long-wavelength-sensitive (M and L) cones. We found that an S-cone grating presented to one eye rivalled with an orthogonal grating presented to the other. Rivalry persisted over a range of luminances and contrasts of the S-cone stimuli, and was greater than could be accounted for by nonrival fading. The spatial spread of rivalry from S-cone stimuli is similar to that for the same stimuli when visible also to the M and L cones (luminance stimuli). We found that an S-cone stimulus would rival with a luminance stimulus, and exploited this to determine the equivalent luminance contrast of S-cone stimuli by putting them in a rivalry competition with luminance stimuli. For rivalry, the equivalent luminance contrast of isoluminant, S-cone stimuli is much less than their S-cone contrast. The existence of rivalry with isoluminant stimuli, along with earlier evidence that such stimuli can support stereopsis, challenges the view that an achromatic channel alone drives certain higher level functions such as depth perception.

Stereoacuity and colour contrast

We have measured the contrast dependence of stereoacuity using both horizontally and vertically oriented, isoluminant (red-green) and isochromatic (yellow-black), 0.5 c/deg Gabor patches. For comparison, contrasts were computed in multiples of detection threshold, where detection threshold was defined as the contrast required for the stimulus to be simultaneously detectable in each eye. Disparity thresholds (1/stereoacuity) for vertical chromatic Gabors were higher than those for vertical luminance Gabors by a factor of between 4 and 9 depending on contrast, and declined less steeply with contrast. Disparity thresholds for horizontal chromatic Gabors were very high (130-210 min arc) compared with horizontal luminance Gabors (by a factor of between 9 and 17) and were only measurable at contrasts above 10 times simultaneous monocular detection threshold. These results support the view that chromatic stereoscopic processing is less precise than luminance stereoscopic processing, and that there is a special deficit in the processing of disparity with horizontally oriented chromatic stimuli. The implications of these results for the role of colour vision in stereopsis are discussed.

Differences between stereopsis with isoluminant and isochromatic stimuli

Contrast thresholds for stereoscopic depth identification (crossed or uncrossed) were measured as a function of disparity by use of isoluminant (red-green) and isochromatic (yellow-black) 0.5 cycles/deg Gabor patches. For the purposes of comparison, stimulus contrasts were scaled by their respective detection thresholds. The Gabor patches could be either vertically or horizontally oriented. It was found that the disparity dependence of the depth-identification contrast thresholds was similar for both chromatic and luminance patterns if the stimuli were vertically oriented, with the overall level of performance worse for the chromatic patterns by a factor of approximately 2 (6 dB). With horizontal patterns this difference was much larger, by a factor of approximately 7 (17 dB). These results suggest first that stereopsis in the absence of luminance cues is supported by a less-contrast-sensitive linear mechanism than that which supports stereopsis in the presence of luminance cues and second that the corresponding nonlinear chromatic stereo mechanism is either nonexistent or very weak. The implications of these data for previous studies of stereopsis at isoluminance is discussed.

The contribution of color to depth perceived from motion parallax

Perceived depth was measured in a colored stimulus while stimulus movement yoked to head displacement simulated a depth of 1 cm. Velocity judgments were also made for similar stimuli moving at the same average speed but without head movement. Both measures decreased to a minimum of about 30-40% of the veridical values when the stimuli were equiluminous. Perceived depth and speed also decreased for a monochromatic stimulus as a function of luminance contrast but much more abruptly than for the chromatic stimuli. The results indicate that equiluminous color stimuli contribute to the perception of depth from motion parallax and that the contribution is not mediated by residual luminance.

Contrast thresholds for stereoscopic depth identification with isoluminant and isochromatic stimuli

Contrast thresholds for stereoscopic depth identification (crossed or uncrossed) were measured as a function of disparity using isoluminant (red-green) and isochromatic (yellow-black) 0.5 c/deg Gabor patches. For the purposes of comparison stimulus contrasts were scaled by their respective detection thresholds. The detection thresholds employed were computed from the monocular detection thresholds of the stereo half-images, based on the assumption that simultaneous detection of these half-images in each eye was a sufficient condition for stereopsis. It was found that the disparity tuning of both chromatic and luminance mechanisms was similar, with a performance peak for a binocular phase disparity of 50-120 deg. However, more contrast was required, relative to detection threshold, for the chromatic patterns to evoke a sensation of stereoscopic depth. These results suggest that stereopsis in the absence of luminance cues is supported by a less-contrast-sensitive analogue of the system that supports stereopsis in the presence of luminance cues. The results are also consistent with there being a lower density of disparity-selective mechanisms in the chromatic pathway. The implications of these data for previous studies of stereopsis at isoluminance is discussed.

Seeing depth in colour: more than just what meets the eyes

Novel binocular depth illusions obtained from two-dimensional colour images are presented. It is demonstrated that the magnitude of these illusions is based on transverse chromatic aberration (TCA), however, the depth obtained cannot be observed unless specific conditions are met even if the TCA is present. Some form of perceptual organization occurring at and/or beyond the binocular fusion site of the brain, is required for some of these effects to occur. An example of a paradoxical finding leading to this conclusion is the observation that under some conditions the same colour can be perceived on separate depth planes while spatially adjacent colours from opposing ends of the visible spectrum (i.e. red and blue or green) can be perceived on the same depth plane simultaneously within the same image. Further, results show that some form of reference plane is required by the brain to use the colour induced disparity, without which, depth cannot be perceived even if the disparity information is present. This phenomenon is spatially tuned for medium to high frequency components and is still detectable under isoluminant conditions which would support the notion that it requires information from the parvocellular pathway. Binocular lustre and rivaldepth are ruled out as being significant factors in the effect. It is argued that this phenomenon represents an instance of global interactive processes induced by TCA while previous studies on chromostereopsis have concentrated on local aspects. Results of the present study may explain why under certain situations depth can be perceived in coloured images and not under other circumstances where TCA is still present.

Efficiency in detection of isoluminant and isochromatic interference fringes

We examined the limitations imposed by neural factors on spatial contrast sensitivity for both isochromatic and isoluminant gratings. We used two strategies to isolate these neural factors. First, we eliminated the effect of blurring by the dioptrics of the eye by using interference fringes. Second, we corrected our data for additional sensitivity losses up to and including the site of photon absorption by applying an ideal-observer analysis described by Geisler [J. Opt. Soc. Am. A 1, 775 (1984)]. Our measurements indicate that the neural visual system modifies the shape of the contrast-sensitivity functions for both isochromatic and isoluminant stimuli at high spatial frequencies. If we assume that the high-spatial-frequency performance of the neural visual system is determined by a low-pass spatial filter followed by additive noise, then the visual system has a spatial bandwidth 1.8 times lower for isoluminant red-green than for isochromatic stimuli. On the other hand, we find no difference in bandwidth or sensitivity of the neural visual system for isoluminant red-green and S-cone-isolated stimuli.

Colour inputs to random-dot stereopsis

Recently it has been claimed by Livingstone and Hubel that, of three anatomically and functionally distinct visual channels (the magnocellular, parvocellular interblob, and blob channels), only the magnocellular channel is involved in the processing of stereoscopic depth. Since the magnocellular system shows little overt colour opponency, the reported loss of the ability to resolve random-dot stereograms defined only by colour contrast seems consistent with this view. However, Julesz observed that reversed-contrast stereograms could be fused if correlated colour information was added. In the present study, 'noise' (non-corresponding) pixels were injected into random-dot stereograms in order to increase fusion time. All six subjects tested were able to achieve stereopsis in less than three minutes when there was only correspondence in colour and not in luminance, and three when luminance contrast was completely reversed. This ability depends on information about the direction of colour contrast, not just the presence of chromatic borders. When luminance and chromatic contrast are defined in terms of signal-to-noise ratios at the photoreceptor mosaic, chromatic information plays at least as important a role in stereopsis as does luminance information, suggesting that the magnocellular channel is not uniquely involved.

Depth, motion, and static-flow perception at metaisoluminant color contrast

Many experiments concerned with the role of color in depth and motion perception have applied isoluminant random-dot stereograms and cinematograms. The poor performance in the absence of luminance contrast has been associated with color-blindness of stereopsis and motion perception (Livingstone, M.S. & Hubel, D.H. (1987) J. Neurosci. 7, 3416-3468). Nevertheless, isoluminant stimuli are not fully accepted as appropriate tools in isolating central mechanisms (Logothetis, N.K., Schiller, P.H., Charles, E.R. & Hurlbert, A.C. (1990) Science 247, 214-217). In our experiments we use a broad luminance range to test whether color can contribute to a given mechanism when luminance contrast is present but has a strong "veto" effect from opposite luminance contrast, a condition we named "metaisoluminance." There is no fusion in stereopsis under polarity reversal, when only luminance information is given, and reversed-phi phenomenon is experienced for motion. As a third "matching" task, we included polarity-reversed random-dot Glass-patterns, which exhibit "static flow" and also show pattern reversal. We found that color can counteract the effects of polarity reversal by restoring stereoscopic fusion and reversed phi motion and does it with increased efficiency as the hue contrast increases. We found no such effect of color in Glass-patterns. Thus, we showed that the visual system for binocular depth and motion perception is not color-blind, although correlated hue information under metaisoluminance does not appear to yield shape perception.

Vision with equiluminant colour contrast: 2. A large-scale technique and observations

A simple technique is described for producing large-scale, tritanopic displays. The technique reproduces the various phenomena of vision with equiluminous-colour contrast that have previously been reported with red/green stimuli. It is, however, much less demanding technically, robust against artifacts, and can be used on large-scale scenes. One advantage of the technique is that a piece of blue filter can be used individually by each observer to compare quickly tritanopic and luminance conditions.

Spatial frequency mechanisms with short-wavelength-sensitive cone inputs

We have estimated the minimum number and frequency tuning of spatial mechanisms with Short-Wavelength-Sensitive (SWS)-cone inputs. This was accomplished by isolating SWS cones with intense, long-wavelength (yellow) adaptation, and measuring threshold elevation functions for short-wavelength, spatially localized test stimuli masked by obliquely oriented, short-wavelength, cosine gratings. Peak spatial frequency of the cosine gratings and test stimuli varied from 0.25 to 2.83 cpd in 0.5 octave steps. Results derived in this manner demonstrate that SWS cones input to at least two orientation selective mechanisms with peak spatial frequencies of approx. 0.7 and 1.4 cpd, respectively. The frequency tuning of the isolated-SWS-cone, spatial frequency mechanisms resemble closely the lowest two mechanisms measured with luminance modulation (i.e. normal viewing conditions where all cones contribute to the response).

The role of colour as a monocular depth cue

Does colour information play a role in the perception of depth? Its input to stereopsis is weak, and it has been suggested that depth from monocular cues, such as texture gradients, is also abolished at isoluminance (colour contrast with no luminance contrast). We first investigated whether depth from texture gradients disappears at isoluminance. The percept remained unaltered. Further experiments revealed that certain colour gradients (at isoluminance) markedly affected the perceived depth. A gradient in saturation (e.g. red-to-grey) was particularly effective, whereas a red-green hue gradient had no effect on perceived slant. We concluded that colour information can be used by the visual system to encode depth, especially in situations where the visual environments is rich in cues which could be used to signal depth in this way.

Color as a source of information in the stereo correspondence process

Although previous research has shown that depth perception is weak for isoluminant stereograms, the possibility remains that color plays an important role in stereopsis when luminance variations are present. To examine this possibility, we measured the relative contribution of chromatic and luminance cues in solving the correspondence problem for ambiguous "wallpaper" stereograms composed of vertical bars. Using an ideal-observer analysis, we found that chromatic cues were used much more efficiently than luminance cues in disambiguating these stereograms when the patterns were presented on a dark background but were used with about equal efficiency when presented on a light background. Another experiment (using the same wallpaper patterns) showed that chromatic and luminance cues were also used with about equal efficiency in a standard stereo detection task. Some of the implications of these results for theories of stereo vision are discussed.

Spatial sensitization in the short wavelength sensitive pathways under dichoptic viewing conditions

Spatial sensitization (Westheimer) functions were measured under conditions that isolated the short wavelength sensitive pathways. The variable diameter pedestal and the test probe were either presented to the same eye (monoptic presentation) or to different eyes (dichoptic presentation). The most significant new finding was that a dichoptically presented, small, blue pedestal caused threshold elevations of about 1-2 long units for an S cone detected probe. However, a large pedestal caused little or no change in threshold. This result contrasts with previous results using white light stimuli, which showed that steadily presented dichoptic pedestals caused little or no threshold change. Furthermore, we show there is little masking when the probe is detected by the isolated M or L cone pathways. These data thus demonstrate a binocular, size dependent interaction revealed only when S cones detect the probe.

Contribution of human short-wave cones to luminance and motion detection

1. Human short-wave S cone signals are important for colour vision and here we examine whether the S cone signals also contribute to motion and luminance. 2. Detection was measured with moving patterns that selectively stimulated S cones-violet sine-wave gratings of 1 cycle deg-1 on an intense yellowish field. For rates up to 12 Hz, detection was governed by non-directional mechanisms, possibly of a chromatic nature, as shown by three findings: moving gratings had to be suprathreshold for their direction to be identified; the threshold ratio of counterphase flickering versus moving gratings was low; and direction-selective adaptation was essentially absent. 3. Evidence for less sensitive, directional mechanisms includes the following: at high velocity, the direction of movement of the violet gratings can be identified just slightly above the detection threshold; directional adaptation was strong with a suprathreshold test pattern; velocity was seen veridically for clearly suprathreshold patterns; and a counterphase flickering test, added in spatial-temporal quadrature phase to a similar suprathreshold mask, had identical detection and direction-identification thresholds. 4. Interactions of long-wave L cone and S cone signals in direction-selective mechanisms were measured with an orange counterphase grating and a violet counterphase test, both flickering at the same rate and presented in spatial quadrature phase on the yellowish adapting field. Direction identification thresholds, measured as a function of the temporal phase of two gratings, demonstrated both that the S cone signal lags considerably behind the L cone signal (an effect that strongly varies with S cone light adaptation), and more strikingly, the S cone signal summates with a negative sign and thus is effectively inverted in direction-selective mechanisms. 5. Quantitatively similar temporal phase functions were obtained with uniform violet and orange flicker when a luminance discrimination criterion was used: thus the S cone signal summates negatively with the L cone signal for both discrimination of luminance flicker and the direction of motion. 6. The temporal phase functions accurately predicted threshold summation for identifying the direction of motion of a pair of violet and orange gratings moving with the same velocity but with different spatial phase offsets. Once the relative temporal phase lag of the S cones was compensated for, there was linear threshold summation for the violet and orange patterns when presented in effective (physiological) spatial antiphase, and clear cancellation when presented in phase. This and related experiments show a linear summation of S, M and L cone signals for direction detection, with the S cones having a negative sign.(ABSTRACT TRUNCATED AT 400 WORDS)

Limits of binocular fusion in the short wave sensitive ("blue") cones

Stereoscopic depth perception is possible when the short wave sensitive (SWS or "Blue") cones are isolated using a yellow adapting field. We have measured the maximum disparity that can be fused (the diplopia threshold) as a function of the separation between pairs of dots or lines. Under all conditions, these diplopia thresholds are the same for the isolated SWS cones as for the entire visual system. In addition, SWS diplopia thresholds vary as a linear function of dot or line separation, so that they exhibit disparity scaling. Further experiments show that disparity scaling is dependent upon the presence of low spatial frequencies in the stimulus and not upon the retinal eccentricity of stimulation. These data indicate that the SWS cones provide information to the disparity processing system through more than one low spatial frequency channel but not through high frequency ones.

Perception of random-dot symmetry and apparent movement at and near isoluminance

There have been conflicting reports on whether apparent movement in random-dot kinematograms is abolished at isoluminance. The present results suggest that it is, provided that dynamic (uncorrelated) surrounds are used, and the subject has to report the shape of the target rather than the presence of movement in an isolated portion of the target. On the other hand, perception of random-dot symmetry is still possible at isoluminance. The reason for this difference appears to be the need for exact-position information in movement but not symmetry perception. Control experiments suggest that the effects are not due to artefacts such as chromatic aberration in the eye.