不同空间频率通道的深度知觉及其质地的影响

我们设计了四种不同质地的随机点立体图对(RDS),并按照Wilson的空间频率通道的频带,分别利用Butterworth滤波器及Gabor函数对之进行了数字滤波处理,然后进行各种左右眼配对观察.得结果如下:1、Gabor函数滤波后图对的匹配结果好于Butterworth滤波器滤波后的图对.2、滤波后图对的匹配效果与图象质地有关.3、不同滤波的图对匹配大多也能形成立体感.4、在体视不同通道图对的匹配中,右眼对图象的高频成分更为敏感.     

空间频率通道中噪声对体视匹配的影响

本工作希望了解噪声在各空间频率通道中对体视的影响.产生从5%至35%每隔5%的七个等级的随机点噪声图象,参照Wilson四通道模型的频率分别用两种不同滤波器对随机点立体图对(RDS)及噪声图象进行滤波,然后在各通道图对中加入不同通道噪声进行匹配观察.得到:1.高频通道的体视抗干扰能力高于低频通道.2.Gabor函数滤波后图对的体视抗干扰能力高于Butterworth滤波器滤波后的,但均低于未滤波的.3.当噪声大于15%时,低频噪声对低频图对及高频噪声对高、低频图对的体视匹配有抑制作用,而低频噪声对高频图对则影响很小.4.未滤波及各通道图对的最大噪声容限均不超过25%.     

立体图对匹配基元的心理物理学实验的探讨

Marr和Poggio等人从视觉信息加工理论中提出零交叉(Zero-Crossing)是视觉早期加工过程中首先得到的最基本的信息。这在生理上有二个高斯分布叠加的感受野以及侧抑现象等可以作为依据,但在心理物理上仍缺乏佐证。我们利用不同密度的随机点,加上了不同百分比的反差噪声的立体图对,观察了各种条件下体视对噪声的耐受特性,得出了有利于说明黑白相间的边(零交叉)为匹配基元的结果。本文报导了这些实验结果并进行了讨论。     

How does the brain sustain a visual percept?

Perception involves the processing of sensory stimuli and their translation into conscious experience. A novel percept can, once synthesized, be maintained or discarded from awareness. We used event-related functional magnetic resonance imaging to separate the neural responses associated with the maintenance of a percept, produced by single-image, random-dot stereograms, from the response evoked at the onset of the percept. The latter was associated with distributed bilateral activation in the posterior thalamus and regions in the occipito-temporal, parietal and frontal cortices. In contrast, sustained perception was associated with activation of the pre-frontal cortex and hippocampus. This observation suggests that sustaining a visual percept involves neuroanatomical systems which are implicated in memory function and which are distinct from those engaged during perceptual synthesis.     

Stereo disparity computation using Gabor filters

A solution to the correspondence problem for stereopsis is proposed using the differences in the complex phase of local spatial frequency components. One-dimensional spatial Gabor filters (Gabor 1946; Marcelja 1980), at different positions and spatial frequencies are convolved with each member of a stereo pair. The difference between the complex phase at corresponding points in the two images is used to find the stereo disparity. Disparity values are combined across spatial frequencies for each image location. Three-dimensional depth maps have been computed from real images under standard lighting conditions, as well as from random-dot stereograms (Julesz 1971). The algorithm can discriminate disparities significantly smaller than the width of a pixel. It is possible that a similar mechanism might be used in the human visual system.     

A scheme for binocular depth perception suggested by neurophysiological evidence

A scheme suggested by neurophysiological evidence is proposed to account for the perceptual phenomena related to binocular stereopsis, especially those observed with Julesz' random stereograms. In the scheme, monocular local features are extracted first. Then the correspondence between the left and right local features is searched for. The correspondence is not one-to-one in general. Thus a sort of direction column due to Blakemore is formed. Each unit in the column is binocular and the receptive field belonging to one eye is located in the same part of the visual field as long as the unit belongs to the same column. However, the receptive fields belonging to the other eye are horizontally displaced to one another. That is, each unit is characterized by binocular disparity. If the correspondence is not one-to-one, then several units belonging to the same column respond simultaneously. Binocular stereopsis can be established if one-to-one correspondence is determined to yield global three dimensional regions. The determination of one-to-one correspondence is carried out through a sort of laterally interacting circuitry in the disparity domain. After the determination of local correspondence, three dimensional global regions are formed by detecting the boundary and by filling-in occluded regions. The results of computer simulation are presented regarding Julesz' stereograms with various types of perturbation. Furthermore, the case of random-dot stereogram in which there is a size difference between the left and right images is simulated. Finally, the computer simulation related to the hysteresis in binocular depth perception is carried out.     

Analysis of a cooperative stereo algorithm

Marr and Poggio (1976) recently described a cooperative algorithm that solves the correspondence problem for stereopsis. This article uses a probabilistic technique to analyze the convergence of that algorithm, and derives the conditions governing the stability of the solution state. The actual results of applying the algorithm to random-dot stereograms are compared with the probabilistic analysis. A satisfactory mathematical analysis of the asymptotic behaviour of the algorithm is possible for a suitable choice of the parameter values and loading rules, and again the actual performance of the algorithm under these conditions is compared with the theoretical predictions. Finally, some problems raised by the analysis of this type of cooperative algorithm are briefly discussed.     

A computer-aided method for the evaluation of fusional reserves with objective control of fusion break

A method for evaluating the fusional reserves has been developed and tested. The method employs an interactive computer program permitting the checking of the subject’s report at the moment of fusion failure during an artificial increase in the convergence or divergence of the visual axes in the course of measurement. The control is based on the use of purely binocular stimuli (random-dot stereograms, RDSs) and dynamic markers, which are added to the test object (TO) in a random order during the measurement procedure and changes in the shape of the TO. The TO with a marker encoded in the stereogram is only perceived owing to binocular mechanisms, and it becomes invisible if fusion has been disturbed. Its disappearance allows the subject to determine the moment of fusion failure, and the shape of the object with the marker before this moment allows the experimenter to check the subject’s report. Statistical analysis of the results of the primary and secondary measurements of the fusional reserves of 72 subjects has confirmed that the method is sufficiently reliable.     

Disparity estimation through Green’s functions of matching equations

Binocular disparities arise from positional differences of scene features projected in the two retinae, and constitute the primary sensory cue for stereo vision. Here we introduce a new computational model for disparity estimation, based on the Green’s function of an image matching equation. When filtering a Gabor-function-modulated signal, the considered Green’s function yields a similarly modulated but shifted version of the original signal. Since a Gabor function models the receptive field of a cortical simple cell, the Green’s kernel thus allows the simulation of relative shifts between the cell’s left and right binocular inputs. A measure of the local degree of matching of such shifted inputs can then be introduced which affords disparity estimation in a similar manner to the energy model of the complex cortical cells. We have therefore effectively reformulated, in physiologically plausible terms, an image matching approach to disparity estimation. Our experiments show that the Green’s function method allows the detection of disparities both from random-dot and real-world stereograms. Partially supported by CNPq-Brazil.     

The science and craft of autostereograms

Autostereograms or SIRDS (Single Image Random-Dot Stereograms) are camouflaged stereograms which combine the Julesz random-dot stereogram principle with the wallpaper effect. They can represent any 3D shape on a single image having a quasi-periodic appearance. Rather large SIRDS can be interpreted in depth with unaided eyes. In the hands of computer graphic designers, SIRDS spread all over the world in 1992-1994, and these images, it was claimed, opened a new era of stereoscopic art. Some scientific, algorithmic and artistic aspects of these images are reviewed here. Scientifically, these images provide interesting cues on stereoscopic memory, and on the roles of monocular regions and texture boundaries in stereopsis. Algorithmically, problems arising with early SIRDS, such as internal texture repeats or ghost images are evoked. Algorithmic recommendations are made for gaining a better control on the construction of SIRDS. Problems of graphic quality (smoothness of the represented surfaces, or elimination of internal texture repeats) are discussed and possible solutions are proposed. Artistically, it is proposed that SIRDS should become less anecdotal, and more oriented towards simple geometric effects, which could be implemented on large panels in natural surrounds.     

A model of binocular stereopsis including a global consistency constraint

 The binocular correspondence problem was solved by implementing the uniqueness constraint and the continuity constraint, as proposed by Marr and Poggio [Marr D, PoggioT (1976) Science 194: 283–287]. However, these constraints are not sufficient to define the proper correspondence uniquely. With these constraints, random-dot stereograms (RDSs), consisting of the periodic textures in each image, are treated as a correspondence of surfaces composed of patches of alternating values of disparity. This is quite different from the surface we perceive through the RDSs, that is a surface characterized by a single depth. Because these constraints are local, they cannot produce the global optimum of correspondence. To obtain the global optimum of correspondence, we propose a model of binocular stereopsis in which a global measure of correspondence is explicitly employed. The model consists of two hierarchical systems. First, the lower system processes various correspondences based on the uniqueness constraint. Second, the higher system provides a global measure of correspondence for the disparity in question. The higher system uniquely determines the global optimum of correspondence in the lower system through the recurrent loop between hierarchical systems. The convergence of the recurrent loop is determined by the consistency between the hierarchical systems. The condition is termed the `global consistency constraint. Received: 27 August 1998 / Accepted in revised form: 8 November 1999     

Pre-attentive segmentation and correspondence in stereo

Traditional stereo grouping models have focused on the problem of stereo correspondence between monocular inputs. Recent physiological data revealed that the disparity selective V2 cells increase their responses when (random-dot stereograms) stimuli within their receptive fields are at or near the boundary of a depth surface. Such highlights to depth (non-luminance) edges are seemingly not computationally required for the correspondence problem. Computationally, these highlights make the boundaries of a depth surface more salient, serving pre-attentive segmentation (between depth planes) and attracting visual attention. In special cases, they enable the psychophysically observed perceptual pop-out of a target from a background of visually identical distractors at a different depth. To achieve the highlights, mutual inhibition between disparity selective cells that are tuned to the same or similar depths is required. However, such mutual inhibition would impede the computation for the correspondence problem, which requires mutual excitation between the same cells. In this work, I introduce a computational model that, I believe, is the first to address both stereo correspondence and pre-attentive stereo segmentation. The computational mechanisms in the model are based on intracortical interactions in V2. I will demonstrate that the model captures the following physiological and psychophysical phenomena: (i) depth-edge highlighting; (ii) disparity capture; (iii) pop-out; and (iv) transparency.     

Human stereovision without localized image features

Many theories of human stereovision are based on feature matching and the related correspondence problem. In this paper, we present psychophysical experiments indicating that localized image features such as Laplacian zerocrossings, intensity extrema, or centroids are not necessary for binocular depth perception. Smooth one-dimensional intensity profiles were combined into stereograms with mirror-symmetric half-images such that these localized image features were either absent or did not carry stereo information. In a discrimination task, subjects were asked to distinguish between stereograms differing only by an exchange of these half-images (ortho- vs. pseudoscopic stereograms). In a depth ordering task, subjects had to judge which of the two versions appeared in front. Subjects are able to solve both tasks even in the absence of the mentioned image features. The performance is compared to various possible stereo mechanisms. We conclude that localized image features and the correspondences between them are not necessary to perceive stereoscopic depth. One mechanism accounting for our data is correlation or mean square difference. Received: 8 February 1994 / Accepted in revised form: 15 September 1994     

Low-level and high-level processes in apparent motion

When a group of dots within a random-dot array is discontinuously displaced, it appears as a moving region perceptually segregated from its stationary surround. The spastial, temporal and other constraints governing this effect are markedly different from those classically found for the apparent motion of isolated stimulus elements. The random-dot display appears to tap a low-level motion-detecting process, distinct from the more interpretive process elicited by the classical displays. The distinct contributions of these processes can be identified in 'multi-stable' displays which yield alternative percepts of apparent motion depending on which one or both of the processes is activated. Such experiments illustrate the interaction of relatively stimulus-constrained and relatively autonomous processes invisual perception.     

Pupils' perceptions of three-dimensional structures in biology lessons

An investigation into 11–15 year old pupils' ability to tackle tasks concerned with the understanding of threedimensional structures studied in school biology lessons is outlined. The investigation takes the form ofapaperand- pencil test, and the test items call for the interpretation of sectional views of three-dimensional structures (represented by diagrams and stereograms) studied in conventional biology syllabuses for 11–15 year olds. Test scores were subject to cluster analysis and the emergent groups are described.

The analysis pointed to the existence of two discrete skills used by pupils interpreting three-dimensional structures. These are the skill of abstracting sectional shapes and that of appreciating the spatial relationships of the internal parts of a structure-. Extracts from tape recordings of pupils ‘talking their way through’ selected tasks corroborate the two-skill theory.     

Neuronal model for stereoscopic vision

A neuronal model for stereopsis is described and simulated. Without the assumption of specific feature detectors, objects are unambiguously located in three-dimensional visual space. Random-dot stereograms are correctly resolved in depth with stimulus details conserved within planar contours.     

Interocular motion combination for dichoptic moving stimuli

When gratings moving in different directions are presented separately to the two eyes, we typically perceive periods of the combination of motion in the two eyes as well as periods of one or the other monocular motions. To investigate whether such interocular motion combination is determined by the intersection-of-constraints (IOC) or vector average mechanism, we recorded both optokinetic nystagmus eye movements (OKN) and perception during dichoptic presentation of moving gratings and random-dot patterns with various differences of interocular motion direction. For moving gratings, OKN alternately tracks not only the direction of the two monocular motions but also the direction of their combined motion. The OKN in the combined motion direction is highly correlated with the perceived direction of combined motion; its velocity complies with the IOC rule rather than the vector average of the dichoptic motion stimuli. For moving random-dot patterns, both OKN and perceived motion alternate only between the directions of the two monocular motions. These results suggest that interocular motion combination in dichoptic gratings is determined by the IOC and depends on their form.     

周期性随机点立体图的原理及设计方法

叙述了一种周期性随机点立体图的原理和设计方法, 并给出一个随机点立体图的实例和程序框图,讨论了周期的大小对深度感知的影响.     

Stereoscopic correspondence for ambiguous targets is affected by elevation and fixation distance

Binocular correspondence must be determined if disparity is to be used to provide information about three-dimensional shape. The current study investigated whether knowledge of the statistical distribution of disparities in the natural environment is employed in this process. A simple model, which produces distributions of distances similar to those found in the natural environment, was used to predict the distribution of disparities in natural images. This model predicts that crossed disparities will be more likely as (i) stimulus elevation decreases below fixation and (ii) fixation distance increases. To determine whether these factors influence binocular correspondence for human observers, ambiguous stereograms were presented to observers, as stimulus elevation and fixation distance were manipulated. Clear biases were observed in the depth perceived in these stereograms, which were more likely to be seen as closer than fixation (i) for stimuli presented below fixation and (ii) as fixation distance increased. These results suggest that binocular correspondence is determined in a manner consistent with the distributions of disparities expected in natural scenes.     

Stages in the construction of stereograms of molecular models

In this article an analysis is performed of the results of stereogram construction using computer programs that model chemical compounds. Considerations about how best to represent models of molecules and improve legibility of stereograms are presented. An original diagrammatic substitute for the picture of a sphere, suitable for application in stereoscopic models of molecules, is proposed.