Preattentive Vision Research Articles

Neglect

Neglect is a disorder of orienting in which patients are unaware of objects in their contralesional visual field. Yet their pre-attentive vision is still able to parse the scene to segregate figure from ground, group objects, and define their primary axis. Therefore, it appears that perceptual processing may be intact up to the level of semantic classification, and that neglect only acts at the level of selection for action and access to awareness. Several mechanisms contribute to neglect, including disinhibited orienting to the ipsilesional field, a deranged representation of space, and deficits in disengaging attention, oculomotor corollary discharge, and representation of contralesional movement trajectories. Recent studies have begun to identify the neural substrates involved in these mechanisms.

Comment on “A Model for Studying Display Methods of Statistical Graphics”

In his opening section Cleveland writes provocatively: is not true that the selection of statistical information is the only critical part and that virtually any method of display suffices. This is so aptly demonstrated that even a reader who only browses the article benefits from looking at the sample graphs. The good/bad examples show how important properties of time curves are disclosed by selecting the right aspect ratio, and how grid lines can greatly facilitate comparisons between displays. Another example shows how data from experimental design with as many as three factors can be shown in a simple two-dimensional display so that one sees the main effects and interactions at work before turning to abstract p values and F ratios based on often doubtful assumptions. Why do some graphs work and others do not? There are scientific reasons for it that have to do with the properties of the human visual system, and these reasons are the focus of the article. However, the goal is not to contribute to the science of visual information processing, but rather, as the author states, it is an engineering one: It is to derive rationales for the design of better display methods. 1. PARALLEL AND SERIAL VISUAL PROCESSES As many authors in vision research, Cleveland feels that there are two profoundly different modes of visual processing. He calls them pattern and table lookup. Pattern perception consists of detection of graphical elements, grouping them so to detect structure and compare quantitative physical aspects of different elements. Table look-up means to actually decode quantitative or categorical information in the data, mainly by relating them to the scales given in the display. This distinction is reminiscent of, but not identical to, a common distinction in vision research: There is much empirical evidence to support a primary, fast, parallel visual process that instantaneously integrates information over the whole visual field that has been called preattentive vision (Julesz 1981; 1984). There is also evidence that supports a secondary visual process that can only process a small part of the visual field at a time, and that therefore scans the field in a slow and serial way. This process has been called

The role of features in preattentive vision: Comparison of orientation, motion and color cues

Arrays of lines or blobs were used to investigate the role of features vs feature contrast in preattentive vision. This study continues earlier work on orientation cues and extends it into the dimensions of motion and color. Tests were performed on patterns displaying continuous feature gradients, i.e. continuous variation, from element to element, in either line orientation, direction of motion, or color. In different series of experiments, the following four aspects of visual perception were investigated: (i) detection of a salient target (“odd man out” paradigm), (ii) segmentation of texture fields, (iii) search strategies for given targets, and (iv) figure-ground discrimination by grouping. Features were, in general, not found to play an important role in these tasks and performance was instead related to feature contrast. Only in the case of color did performance also depend upon hue, i.e. feature properties themselves. Whereas in all dimensions tested pop-out (“saliency”) and segmentation were obtained from target or border elements whose local feature contrast was well above the level of variation elsewhere in the pattern, performance in the two latter tasks differed between orientation, motion, and color. In search, orientation targets were detected quickly when sufficiently distinct from their neighbors, but were apparently searched for serially if feature contrast was similar to that of other elements nearby. Color targets, however, were always detected fast in these patterns, independently of local feature contrast. Also, the perceived grouping of orientation or motion defined targets depended only on local feature contrast and not on the similarity of target elements. In fact, figures of dissimilar elements were seen as easily as figures of similar elements, indicating that, in these dimensions, stimulus coherence is not essential for the discrimination of figure and ground. For color, however, figures made up of the same targets were always seen slightly better than figures composed of different colors.

Talking to yourself aboutwhat is where: What is the vocabulary of preattentive vision?

Talking to yourself about<i>what is where</i>: What is the vocabulary of preattentive vision?

Identifying High Level Features of Texture Perception

A fundamental issue in texture analysis is that of deciding what textural features are important in texture perception, and how they are used. Experiments on human preattentive vision have identified several low-level features (such as orientation of blobs and size of line segments), which are used in texture perception. However, the question of what higher level features of texture are used has not been adequately addressed. We designed an experiment to help identify the relevant higher order features of texture perceived by humans. We used 20 subjects, who were asked to perform an unsupervised classification of 30 pictures from Brodatz′s album on texture. Each subject was asked to group these pictures into as many classes as desired. Both hierarchical cluster analysis and nonparametric multidimensional scaling (MDS) were applied to the pooled similarity matrix generated from the subjects′ groupings. A surprising outcome is that the MDS solutions fit the data very well. The stress in the two-dimensional case is 0.10, and the stress in the three-dimensional case is 0.045. We rendered the original textures in these coordinate systems, and interpreted the (rotated) axes. It appears that the axes in the 2D case correspond to periodicity versus irregularity, and directionality versus nondirectionality. In the 3D case, the third dimension represents the structural complexity of the texture. Furthermore, the clusters identified by the hierarchical cluster analysis remain virtually intact in the MDS solution. The results of our experiment indicate that people use three high-level features for texture perception. Future studies are needed to determine the appropriateness of these high-level features for computational texture analysis and classification.

Feature analysis and the role of similarity in preattentive vision.

Texture arrays of line elements at various orientations were used to study three phenomena of preattentive vision. Subjects were asked (1) to discriminate texture areas and to distinguish their form (experiments on texture segmentation); (2) to detect salient or vertical line elements (experiments on pop-out); and (3) to identify configurations of similar or or dissimilar targets (experiments on grouping). Within the patterns, line orientation was systematically varied to distinguish the effect of differences between areas from the effect of similarity within areas. In all of the experiments, performance was found to depend on local orientation contrast at texture borders rather than on the analysis of line orientation itself. Texture areas were correctly identified only when the orientation contrast at the border well exceeded the overall variation of line orientation in the pattern. Similarly, only target elements with high local orientation contrast were detected fast and "in parallel". Targets with an orientation contrast lower than background variation required serial search. Preattentive grouping was found to depend on saliency, as defined by local orientation contrast, but not on the similarity of line elements. In addition to local orientation contrast, which played an important role in all of the visual phenomena studied, influences from the alignment of line elements with the outline of a figure were also seen.

Visual search for direction of shading is influenced by apparent depth.

Recent reports of rapid visual search for some feature conjunctions suggested that preattentive vision might be sensitive to scene-based as well as to image-based features (Enns & Rensink, 1990a, 1990b). This study examined visual search for targets defined by the direction of a luminance gradient, a conjunction of luminance and relative location that often corresponds to object curvature and direction of lighting in naturalistic scenes. Experiment 1 showed that such search is influenced by several factors, including the type of gradient, the shape of the contour enclosing the gradient, and the background luminance. These factors were varied systematically in Experiment 2 in a three-dimensionality rating task and in a visual-search task. The factors combined interactively in the rating task, supporting the presence of an emergent property of three-dimensionality. In contrast, each factor contributed only additively to the speed of the visual-search task. This is inconsistent with the view that search is guided by specialized detectors for surface curvature or direction of lighting. Rather, it is in keeping with the view that search is governed by a number of "quick and dirty" processes that are implemented rapidly and in parallel across the visual field.

The role of the superior colliculus in facilitating visual attention and form perception.

When interhemispheric transfer in cats is studied from an intact hemisphere to a hemisphere with a suprasylvian cortical lesion, excellent transfer of grating discriminations, but no transfer of forms, is present. Stimuli with global, repetitive features covering a large visual field (gratings), which can be discriminated by preattentive vision, are transferred; perception of stimuli with local features (forms), which require serial exploration using focal vision, is defective in the hemisphere with cortical lesion and transfer is lacking. Influence of the midbrain in facilitating focal vision is shown by the restoration of form discriminations after section of the superior collicular commissure. It is hypothesized that the perceptual defect after lesion in the suprasylvian cortex is due to poor spatial attention and its restoration after midbrain lesion is due to improved function of those collicular cells that mediate orienting of attention.

Preattentive Vision and Perceptual Groups

Recent evidence suggests that preattentive processing may not be limited to the analysis of simple stimulus features as previously suggested. To explore this idea a visual search task was used to test whether the shapes of several perceptual groups can be processed in parallel. Textured displays that give rise to strong perceptual grouping were used to create figures on a background. Search times for a target figure distinguished by a unique shape were found to be independent of the number of distractor figures in the display. This result indicates that perceptual groups may be processed in parallel and suggests an expanded role for preattentive processing in vision.

Stereo and Motion Cues in Preattentive Vision Processing—Some Experiments with Random-Dot Stereographic Image Sequences

Low-level preattentive vision processing is of special interest since it seems the logical starting point of all vision processing. Exploration of the human visual processing system at this level is, however, extremely difficult, but can be facilitated by the use of stroboscopic presentation of sequences of random-dot stereograms, which contain only local spatial and temporal information and therefore limit the processing of these images to the low level. Four experiments are described in which such sequences were used to explore the relationships between various cues (optical flow, stereo disparity, and accretion and deletion of image points) at the low level. To study these relationships in more depth, especially the resolution of conflicting information among the cues, some of the image sequences presented information not usually encountered in 'natural' scenes. The results indicate that the processing of these cues is undertaken as a set of cooperative processes.

The medium is not the message in preattentive vision.

The medium is not the message in preattentive vision.

A neural network architecture for preattentive vision.

Recent results towards development of a neural network architecture for general-purpose preattentive vision are summarized. The architecture contains two parallel subsystems, the boundary contour system (BCS) and the feature contour system (FCS), which interact together to generate a representation of form-and-color-and-depth. Emergent boundary segmentation within the BCS and featural filling-in within the FCS are herein emphasized within a monocular setting. Applications to the analysis of boundaries, textures, and smooth surfaces are described, as is a model for invariant brightness perception under variable illumination conditions. The theory shows how suitably defined parallel and hierarchical interactions overcome computational uncertainties that necessarily exist at early processing stages. Some of the psychophysical and neurophysiological data supporting the theory's predictions are mentioned.

Developing a quantitative model of human preattentive vision

To develop robust computer vision system methodologies, ones that have a range of applicability, early vision operators capable of matching a level of human perceptual performance are required. This implies the need to develop operators that can perform a variety of image analysis tasks in a unified and consistent fashion. These image analysis tasks include finding boundaries between regions of uniform but different gray levels and textures. They also include gauging Gestalt grouping concepts such as uniformity and proximity, so that these concepts are incorporated into the segmentation process. Lastly, the operators should be able to gauge information that allows the characteristics of surfaces to be made explicit, e.g., so-called shape-from-shading and shape-from-texture methods. Developing robust methods for performing these tasks corresponds to developing a quantitative model of human preattentive vision. A basis for such a quantitative model is proposed that is contrary to current theories in computer vision but that seemingly provides a unified method for image analysis tasks and to explain a number of perceptual phenomena. >

Seeing textons in context.

Texton theory holds that visual texture segregation occurs through the preattentive detection of local differences in primitive visual units called “textons.” Three textons have been proposed for shape: lines, line terminators, and line intersections. The experiments reported here show that line closure also behaves like a texton under some conditions. In addition, the experiments show that texture segregation is not determined by texton differences per se, but by the extent to which the unique textons in a region are salient in the context of the textons common to regions. This suggests that preattentive vision is not as simple minded as texton theory claims it to be.

Conjunction of color and form without attention: Evidence from an orientation-contingent color aftereffect.

According to feature-integration theory (Treisman & Gelade, 1980), separable features such as color and shape exist in separate maps in preattentive vision and can be integrated only through the use of spatial attention. Many perceptual aftereffects, however, which are also assumed to reflect the features available in preattentive vision, are sensitive to conjunctions of features. One possible resolution of these views holds that adaptation to conjunctions depends on spatial attention. We tested this proposition by presenting observers with gratings varying in color and orientation. The resulting McCollough aftereffects were independent of whether the adaptation stimuli were presented inside or outside of the focus of spatial attention. Therefore, color and shape appear to be conjoined preattentively, when perceptual aftereffects are used as the measure. These same stimuli, however, appeared to be separable in two additional experiments that required observers to search for gratings of a specified color and orientation. These results show that different experimental procedures may be tapping into different stages of preattentive vision.

Direction- and velocity-specific responses from beyond the classical receptive field in the middle temporal visual area (MT).

The true receptive field of more than 90% of neurons in the middle temporal visual area (MT) extends well beyond the classical receptive field (crf), as mapped with conventional bar or spot stimuli, and includes a surrounding region that is 50 to 100 times the area of the crf. These extensive surrounds are demonstrated by simultaneously stimulating the crf and the surround with moving stimuli. The surrounds commonly have directional and velocity-selective influences that are antagonistic to the response from the crf. The crfs of MT neurons are organized in a topographic representation of the visual field. Thus MT neurons are embedded in an orderly visuotopic array, but are capable of integrating local stimulus conditions within a global context. The extensive surrounds of MT neurons may be involved in figure-ground discrimination, preattentive vision, perceptual constancies, and depth perception through motion cues.

Human Factors and Behavioral Science: Textons, The Fundamental Elements in Preattentive Vision and Perception of Textures

Recent research in texture discrimination has revealed the existence of a separate “preattentive visual system” that cannot process complex forms, yet can, almost instantaneously, without effort or scrutiny, detect differences in a few local conspicuous features, regardless of where they occur. These features, called “textons”, are elongated blobs (e.g., rectangles, ellipses, or line segments) with specific properties, including color, angular orientation, width, length, binocular and movement disparity, and flicker rate. The ends-of-lines (terminators) and crossings of line segments are also textons. Only differences in the textons or in their density (or number) can be preattentively detected while the positional relationship between neighboring textons passes unnoticed. This kind of positional information is the essence of form perception, and can be extracted only by a time-consuming and spatially restricted process that we call “focal attention”. The aperture of focal attention can be very narrow, even restricted to a minute portion of the fovea, and shifting its locus requires about 50 ms. Thus preattentive vision serves as an “early warning system” by pointing out those loci of texton differences that should be attended to. According to this theory, at any given instant the visual information intake is relatively modest.