Hirohisa Yaguchi

Motion in depth based on inter-ocular velocity differences.

Two different binocular cues are known for detecting motion in depth. One is disparity change in time and the other is inter-ocular velocity difference. In contrast to the well known fact of the use of the disparity cues, no evidence of contribution of inter-ocular velocity differences for detecting motion in depth has been reported. We demonstrate that motion in depth can be seen based solely on inter-ocular velocity differences using binocularly uncorrelated random-dot kinematograms. This indicates that the visual system uses monocular velocity signals for processing motion in depth in addition to disparity change in time.

Subadditivity and superadditivity in heterochromatic brightness matching

Heterochromatic brightness matches were carried out between bichromatic mixtures of lambda 1 and lambda 2, and a 100 td white reference light of 2 degrees arc. Two characteristic properties of additivity failure were observed. One is the asymmetrical property of additivity failure, that is a remarkable brightness reduction at the luminance ratio where the hue cancellation occurred. The other is the superadditivity for a mixture of a violet and a white or a middle-wave light. These properties were accounted for by a nonlinear vector model on the assumption that the opponent-color channels outputs were nonlinearly transferred to the brightness perception.

Brightness luminous-efficiency functions for 2° and 10° fields

CIE Vλ is not representative of luminous-efficiency function based on heterochromatic brightness matching. CIE Technical Committee 1.4 Vision (TC 1.4) presented a 2° brightness-matching luminous-efficiency function based on studies of a total of 31 observers to supplement the Vλ. In view of their importance to illuminating engineering and physiological optics, we analyzed the various conditions under which these seven studies were conducted. Data from three of the groups are considered inappropriate, and we revised the TC 1.4 brightness luminous-efficiency function based on the remaining 19 subjects. Data from 18 Japanese subjects, coming from five research groups, are added to the above subjects, and an averaged luminous-efficiency function is derived. The result does not appreciably differ from the revised TC 1.4 function and is considered to represent a brightness-matching standard luminous-efficiency function for a 2° field. A brightness luminous-efficiency function for a 10° field based on nine Japanese subjects is presented. It differs from the 2° function only at short wavelengths when the functions are normalized at 570 nm. A theoretical approach for using the brightness-matching luminous-efficiency function to assess the brightness of 2° broadband sources is introduced, and some numerical examples are given.

Detection of relative and uniform motion

We measured the lowest velocity (velocity threshold) for discriminating motion direction in relative and uniform motion stimuli, varying the contrast and the spatial frequency of the stimulus gratings. The results showed significant differences in the effects of contrast and spatial frequency on the threshold, as well as on the absolute threshold level between the two motion conditions, except when the contrast was 1% or lower. Little effect of spatial frequency was found for uniform motion, whereas a bandpass property with a peak at approximately 5 cycles per degree was found for relative motion. It was also found that contrast had little effect on uniform motion, whereas the threshold decreased with increases in contrast up to 85% for relative motion. These differences cannot be attributed to possible differences in eye movements between the relative and the uniform motion conditions, because the spatial-frequency characteristics differed in the two conditions even when the presentation duration was short enough to prevent eye movements. The differences also cannot be attributed to detecting positional changes, because the velocity threshold was not determined by the total distance of the stimulus movements. These results suggest that there are two different motion pathways: one that specializes in relative motion and one that specializes in uniform or global motion. A simulation showed that the difference in the response functions of the two possible pathways accounts for the differences in the spatial-frequency and contrast dependency of the velocity threshold.

A new method for assessing motion-in-depth perception in strabismic patients

Aim: In strabismus clinics, stereoscopic depth perception is usually examined using static stimuli, but these stimuli do not necessarily allow assessment of the ability to perceive motion in depth. We assessed the ability to perceive motion-in-depth perception using a novel stereo motion test that we developed and compared with that to perceive static depth perception using a conventional stereo test in strasbismic patients. Methods: To investigate motion-in-depth perception in patients with strabismus, we developed a stereo motion test using four types of computer-generated dynamic visual stimuli. Three of them are random dot stereograms of two parallel planes moving in depth. The patient is asked to indicate the planes’ direction of rotation in depth (in the first and second types) or the presence/absence of motion-in-depth signal (in the third type). The fourth type of stimulus was a random dot stereogram of a rotating cylinder. The upper and lower parts of the cylinder rotate in opposite directions, and the patient is asked to indicate the position of the border between the two parts. Threshold disparity was defined as the disparity (relative disparity between the nearest and farthest points of the planes or the cylinder) that gives a critical level of performance with the method of limit. The conventional Titmus stereo test using static visual stimuli was used to assess static depth perception. The measurements were performed in 52 strabismic patients, aged between 4 and 38 years old, who visited Tokyo Medical University Hospital between January 2003 and July 2004. Results: The results showed a poor correlation in the threshold of individual patients between the stereo motion test and conventional Titmus stereo test. For example, the ability to perceive motion in depth (disparity threshold <500 sec of arc) was demonstrated in three of seven patients who were not able to perceive depth using static stimuli (0/9 for Titmus circle). These results suggest that the process of the dynamic element of binocular depth perception is preserved in some of the strabismic patients who lack static stereopsis. Conclusion: This study indicates the importance of testing motion-in-depth perception as well as static depth perception in assessing stereopsis in strabismic patients.

Smooth shifts of visual attention

To examine whether visual attention shifts continuously across the visual field, we measured sensitivity to a small flash presented at various locations while the observer was tracking a moving target in an ambiguous apparent motion display. The sensitivity peaked near the target and the peak shifted smoothly along the apparent motion path. Since the peak-shift speed varied with the speed of the tracked target, we conclude that the attention mechanism selects the location to facilitate processing by tracking the target disk continuously. Attention does not simply select a location for enhanced processing, but rather predicts the future location of the object of interest based on its velocity.

Tracking the apparent location of targets in interpolated motion

Under appropriate conditions, a target moving in discrete steps can appear to move smoothly and continuously even within the portions of the path where no physical stimulus is present. We investigated the nature of this interpolated motion in attentive tracking displays as well as apparent motion. The results showed that the apparent location of the target moved smoothly through space between the two discrete locations and the judgements of interpolated motion for attentive tracking and apparent motion were comparable to those for continuous motion in both the perceived path and the precision of the judgements. There were few, if any, differences between judgements for real and interpolated motion. An alignment procedure showed that the smooth change in location judgements was real and not a consequence of averaging across discrete locations actually seen on each trial. We also found that the slowest alternation rate which supported accurate location judgements corresponded to a critical SOA of about 500 ms, similar to the longest SOA which supported a subjective impression of motion in the display. Deviations from a constant velocity which were shorter than 200 ms did not register in the judged motion path, suggesting a fairly long time constant for the integration of velocity information into the perceived motion. These results suggest a specialized motion analysis which provides an accurate, explicit model of the interpolated motion path.

Mesopic luminous-efficiency functions for various adapting levels

The luminous-efficiency functions for a centrally viewed 10 degrees field were measured by heterochromatic brightness matching for various retinal illuminance levels of a reference field at various adapting levels. The subject was always presented with a 45 degrees white adapting light except when the test field for brightness matching was substituted for the adapting field for 500 msec. In order to investigate the contributions of rods and cones to the brightness sensation, the luminous-efficiency curves obtained from two subjects were analyzed with the Ikeda-Shimozono formula. When the subject was presented with an adapting light above about 100 trolands (Td), the luminous-efficiency function became photopic at any luminance level of the test field; similarly, when the test field was at 100 Td, the luminous-efficiency function was photopic at any adapting level.

Individual differences of the contribution of chromatic channels to brightness

Perceived brightness is considered to be a combined consequence of outputs of the luminance channel and the chromatic channels in the visual system. The differences of logarithmic spectral luminous efficiencies between heterochromatic brightness matching and flicker photometry that were obtained from 16 subjects were examined by using principal component analysis. The luminous-efficiency difference between the two methods is described by only two principal components. Individual characteristics of the contribution of chromatic channels to brightness can be specified by measuring luminous efficiencies at 470 and 660 nm.

Differences in temporal frequency tuning between the two binocular mechanisms for seeing motion in depth

There are two types of binocular cues available for perception of motion in depth. One is the binocular disparity change in time and the other is the velocity difference between the left and the right retinal images (inter-ocular velocity differences). We measured the luminance contrast threshold for seeing motion in depth while isolating either of the cues at various temporal modulations of velocity in the stimulus. To isolate disparity cues, dynamic random-dot stereograms were used (the disparity condition) while binocularly uncorrelated random-dot kinematograms were used to isolate velocity cues (the velocity condition). Results showed that sensitivity peaked at a temporal frequency (approximately 1 cps) in the velocity condition while the peak in the disparity condition was at the lowest frequency (0.35 cps) or at least at a frequency lower than that in the velocity condition. This suggests that the visual system has different temporal frequency properties for the velocity and disparity cues for motion in depth.