Human perception

Werkhoven, P., Snippe H.P. & Koenderink J.J. (1990). Effects of element orientation on apparent motion perception. Perception & Psychophysics 47(6), 509‑525, 1990.  We present an ambiguous motion paradigm that allows us to quantify the influence of aspects of form relevant to the perception of apparent motion. We report on the role of bar element orientation in motion paths. The effect of orientation differences between bar elements  in a motion path is small with respect tot the crucial role of the orientation of bar elements relative to motion direction. Motion percpetion between elements oriented along the motion direction dominates motion perception between elements oriented perpendicularly to motion direction. The perception of apparent motion is affected by bar length and width and is anisotropuic.

Werkhoven P., Snippe H.P. & Koenderink J.J. (1990). Metrics for the strength of low level motion perception. J. of Visual Communication and Image Representation 1(2), 176‑188.   It is tempting to explain apparent motion perception in terms of a similarity metric: the perception of apparent motion between two image elements, displayed at different positions in space and different moments in time, is carried by the similarity of the elements. Alternatively, apparent motion perception can be modeled in terms of a covariance metric: motion strength is determined by the product of receptive subunit outputs that sense certain local image qualities (e.g. incremental flux). We present an ambiguous motion paradigm that (in contrast to previously used paradigms) allows to discriminate between a similarity metric and a covariance metric. The results provide clear evidence against a similarity metric. It is shown that motion between elements that differ in some qualities can easily dominate motion between identical elements (which is predicted by a covariance metric). The element qualities examined are contrast, area and incremental flux. This study is aimed at the low level motion domain.

Werkhoven P. & Koenderink J.J. (1990). Interference in rotary motion. J. of the Optical Society America A 7(9), 1627‑1631.   Vorticity discrimination for a rotating annulus is affected by the presence of an interfering randomly rotating concentric annulus, whenever the radii differ less than 20 %. This interaction is interpreted in terms of an activity of local motion detectors that scale linearly in size with the tuning velocity. An interfering annulus decreases the perceived vorticity of the reference-annulus, independent of the vorticity of the interfering annulus.

Werkhoven P. & Koenderink J.J. (1991). Visual processing of rotary motion. Perception & Psychophysics 49(1), 73‑82.  Local descriptions of velocity fields (e.g. rotation, divergence and deformation) contain a wealth of information for form-perception and ego-motion. In spite of this, human psychophysical performance in estimating these entities has not been thoroughly examined yet. In this paper, we report on the visual discrimination of rotary motion. A sequence of image-frames is used to elicit an apparent rotation of an annulus composed of dots in the fronto-parallel plane around a fixation spot at the center of the annulus. Differential angular velocity thresholds are measured as a function of angular velocity, diameter of the annulus, number of dots, display-time per frame and the number of frames. The results show a U-shaped dependence of angular velocity discrimination on spatial scale, with minimal Weber-fractions of 7%. Experiments with a scatter in the distance of the individual dots to the center of rotation demonstrate that angular velocity can not be assessed directly; perceived angular velocity depends strongly on the distance of the dots relative to the center of rotation. We suggest that the estimation of rotary motion is mediated by local estimations of linear velocity.

Werkhoven P. & Koenderink J.J. (1991). Reversed rotary motion perception. J. of the Optical Society America A 8(9), 1510‑1516.  A stroboscopically presented revolving annulus composed of dots is used to elicit rotary motion perception. Observers judge the direction of rotary motion. We find sharp and gradual transitions in the probability for reversed motion perception as a function of the angle of rotation between successive frames. These transitions reveal that matches between non-successive frames can dominate motion perception. The transitions are scale invariant. The strength of a match is discussed in terms of a ``motion strength function'' which is a separable function of the angle of rotation between successive frames and the frame repetition rate. The dependence of motion strength on the frame repetition rate (``time function'') is computed from the transitions. The similarity of this time function for rotary motion with the time function found for linear motion suggests that mechanisms for the discrimination of rotary motion address local detectors of linear motion.

Werkhoven P., Snippe H.P. & Toet A. (1992). Visual processing of optic acceleration. Vision Research 32(23), 2313‑2329.   We examined human sensitivity for temporal modulations We determine human modulation detection thresholds for temporal modulations of speed and motion direction of moving dots. Thresholds, as a function of modulation frequency, all show a temporal low-pass nature. This argues for a temporally smoothed version of the physical velocity signal as the decision variable used by our observers. This obligatory time smoothing is estimated to extend over 100-140~ms (somewhat dependent on stimulus speed) under our present conditions.  Finally we show evidence for a variance detection on this filtered velocity signal as the final detection process used by our observers. We estimate a standard deviation of 17% for the (filtered) speed signal, and a standard deviation of 8% for the (filtered) motion direction signal to be necessary for velocity modulation detection at threshold level to occur.

Werkhoven P. & Koenderink J.J. (1993). Visual size invariance does not apply to geometric angle and vorticity. Perception 22, 177‑184. We examine the ability of humans to visually estimate geometric angle and vorticity (rotational speed), when the spatial scale of the stimuli is varied. Both properties are objectively invariant at different spatial scales. Perhaps surprisingly, our experiments show that the judgement of acute geometric angles as well as the judgement of vorticity varies strongly and monotonically with the scale of presentation. If the image is magnified, the perceived geometric angle and the perceived vorticity increase. If the image is minified, they decrease. This result imposes strong constraints on perceptual theories.

Werkhoven P., Chubb C. & Sperling G. (1993). The dimensionality of texture‑defined motion: A single channel theory. Vision Research 33(4), 463‑485.  What determines the strength of apparent motion perception when the stimulus has no net directional energy in the Fourier domain (texture defined motion)? In a previous paper (Werkhoven et al., 1991) we have demonstrated the counterintuitive finding that the correspondence in spatial frequency and contrast between neighboring patches of texture in a spatiotemporal motion path are not relevant to motion strength. Instead, we found strong support for what we call a single channel or {\it one dimensional} motion computation: a {\it single} nonlinear transformation of the image, followed by standard motion analysis. In this paper, we further studied the dimensionality of the motion computation for a parameter space that includes texture orientation and temporal frequency, in addition to textural properties like spatial frequency and contrast. We used ambiguous motion displays in which one motion path, consisting of patches of nonsimilar texture, competes with one other motion path, having patches of similar texture only. The data show that motion between dissimilar patches of texture (which are orthogonally oriented, have a two octave difference in spatial frequency and differ 50% in contrast) can easily dominate motion between similar patches of texture. An analysis of the data shows that the motion computation is roughly one dimensional for the parameter space examined and invariant for different temporal between 1 and 4 Hz.

Snippe H.P. & Werkhoven P. (1993). Pulse modulation detection in human motion vision. Vision Research 33(5/6), 647‑656.   We present data on the human sensitivity to temporal pulse modulations of target velocity. We measured threshold detection modulation amplitudes for pulse-shaped speed modulations, as a function of pulse duration and temporal frequency.  At short pulse durations (up to 50 ms) and low modulation frequency (1 Hz), detection amplitudes are ruled by a Bloch law: the product of pulse duration and threshold modulation amplitude is a constant. This constant corresponds to a position modulation with an amplitude of 3 arcmin in a coordinate frame that moves at the average speed (3 deg/s) of the target. At longer pulse durations we find deviations from Bloch's law. Speed modulation thresholds are not critically dependent on target luminance contrast. These results are modeled by a modulation detection process in two stages. A functional description of the first stage is filtering of the true speed modulation signal by a second order low-pass filter with a characteristic time constant of 20-25ms. The second (decision) stage is variance detection: modulations are detected when the variance of the filtered modulation function exceeds a certain threshold variance. The square root threshold variance is estimated 8-10%.  This two-parameter model accurately predicts the measured dependence of pulse modulation detection thresholds on pulse duration and pulse density.

Werkhoven P., Sperling G. & Chubb C. (1994). Motion perception between dissimilar gratings: Spatiotemporal properties. Vision Research 34(20), 2741‑2759.   We examine apparent motion carried by textural properties. The texture stimuli consist of patches of sinusoidal grating of various spatial frequencies and contrasts. Phases are randomized between frames to insure that motion mechanisms sensitive to correspondences in stimulus luminance are not systematically engaged. We use ambiguous apparent motion displays in which a heterogeneous motion path defined by alternating patches of texture s (standard) and texture v (variable) competes with a homogeneous motion path defined solely by patches of texture s. Our results support a model in which strength of texture-defined motion is computed from a single spatial transformation of the stimulus - the activity transformation. The value assigned a point in space-time by the activity transformation is directly proportional to local texture contrast and inversely proportional to local spatial frequency (within the range of spatial frequencies examined). Thus, the activity transformation can be modeled as the rectified output of a low-pass spatial filter applied to stimulus contrast. The strength of texture-defined motion between a patch of texture s and a patch of texture v is proportional to the product of the activities of s and v. A counterintuitive implication of this model borne out in our data is that apparent motion along a heterogeneous path consisting of alternating patches of a low contrast, low frequency texture (texture l) and patches of high contrast, high frequency texture (texture h) can be stronger than motion along a homogeneous path of identical patches of texture.

Vries S.C. de & Werkhoven P. (1995). Cross‑modal slant and curvature matching of stereo‑ and motion‑specified surfaces. Perception & Psychophysics, 57(8), 1175-1186.

Werkhoven P. & Veen H.A.H.C. van (1995). Extraction of relief from visual motion. Perception & Psychophysics 57(5), 645-656. We quantified the ability of human subjects to discriminate the relative distance of two points from a slanted plane when viewing the projected velocities of this scene (orthographic projection). The relative distance from a plane (called relief) is a 3D property that is invariant under linear (affine) transformations. As such, relief can {\em in principle} be extracted from the instantaneous projected velocity field; a metric representation, which requires the extraction of visual acceleration, is not required. The stimulus consisted of a slanted plane P (specified by three points) and two points Q1 and Q2 that are non-coplanar with P. This configuration of points oscillated rigidly around the vertical axis. We have measured the {\em systematic error} and {\em accuracy} with which human subjects estimate the relative distance of points Q1 and Q2 from plane P as a function of the slant of P. The systematic error varies with slant: it is low for small slant values, reaches a maximum for medium slant values and drops again for high slant values. The accuracy covaries with the systematic error and is thus high for small and large slant values and low for medium slant values. These results are successfully modeled by a simple relief-from-motion computation based on local estimates of projected velocities. The data are well predicted by assuming (1) a measurement error in velocity estimation that varies proportional to velocity (Weber's law) and (2) an eccentricity dependent underestimation of velocity.

Veen H.A.H.C. van, Kappers A.M.L., Koenderink J.J. & Werkhoven P. (1996). Discriminating the volume of motion-defined solids. Perception & Psychophysics, 58(4), 561‑570.

Veen H.A.H.C. van & Werkhoven P. (1996). Metamerisms in structure-from-motion perception. Vision Research, 36(14), 2197-2210.

Erp, J.B.F. van, Werkhoven, P.J. (2004). Perception of vibro-tactile asynchronies. Perception, 33, pp. 103-111. This study investigated the consistency between tactually and visually designated empty time intervals. In a forced-choice discrimination task participants judged whether the second of two intervals was shorter or longer than the first interval. Two pulses defined the intervals. The pulse was either a vibro-tactile burst presented to the fingertip or a foveally presented white square. The comparisons were made for uni-modal and cross-modal intervals. We used four levels of standard interval durations in the range of 100 - 800 ms. The results showed that tactile empty intervals must be 8.5% shorter to be perceived as long as visual intervals. This cross-modal bias is larger for small intervals and decreases with increasing standard intervals. The Weber fractions (the threshold divided by the standard interval) are 20% and are constant over the standard intervals. This indicates that the Weber law holds for the range of interval lengths tested. Furthermore, the Weber fractions are consistent over uni- and cross modal comparisons, which indicates that there is no additional noise involved in the cross-modal comparisons.