Figure-ground Segregation Research Articles

Attention and figure-ground segregation in a model of motion perception

Attention and figure-ground segregation in a model of motion perception

The problem of the perception of holes and figure-ground segregation in the watercolor illusion

The problem of the perception of holes and figure-ground segregation in the watercolor illusion

Figure-ground segregation and spatial phase tuning of extra-receptive field of V1 neurons in awake monkey

Figure-ground segregation and spatial phase tuning of extra-receptive field of V1 neurons in awake monkey

Masking interrupts figure-ground signals in V1

In a backward masking paradigm, a target stimulus is rapidly (<100 msec) followed by a second stimulus. This typically results in a dramatic decrease in the visibility of the target stimulus. It has been shown that masking reduces responses in V1. It is not known, however, which process in V1 is affected by the mask. In the past, we have shown that in V1, modulations of neural activity that are specifically related to figure–ground segregation can be recorded. Here, we recorded from awake macaque monkeys, engaged in a task where they had to detect figures from background in a pattern backward masking paradigm. We show that the V1 figure–ground signals are selectively and fully suppressed at target–mask intervals that psychophysically result in the target being invisible. Initial response transients, signalling the features that make up the scene, are not affected. As figure–ground modulations depend on feedback from extrastriate areas, these results suggest that masking selectively interrupts the recurrent interactions between V1 and higher visual areas.

Shape-specific perceptual learning in a figure-ground segregation task

Shape-specific perceptual learning in a figure-ground segregation task

The notion of 'purely time-based figure-ground segregation' is still justified

The notion of 'purely time-based figure-ground segregation' is still justified

Areas involved in figure-ground segregation based on luminance, colour, motion, and stereoscopic depth visualized with fMRI

Areas involved in figure-ground segregation based on luminance, colour, motion, and stereoscopic depth visualized with fMRI

Object-sensitive activity reflects earlier perceptual and later cognitive processing of visual objects between 95 and 500 ms

Object-sensitive areas have been defined using functional magnetic resonance imaging (fMRI), but the time course of this cortical activation is largely unknown. In a high-density, event-related brain potential (ERP) version of a prior fMRI study of object-sensitive areas, people categorized intact grayscale pictures of known objects and discriminated them from scrambled versions created by randomizing the phases of the spatial frequency spectrum; no object or parts can be discerned in scrambled versions. Both indirect functional and direct anatomical approaches were taken to integrate ERP and fMRI results. The two-state interactive account of visual object cognition predicts object-sensitivity (intact vs. scrambled) both before and after 200 ms during early and later ERPs that index processes in posterior cortex critical to visual object perception and cognition. As predicted, findings from 57 participants revealed early object-sensitive activation during the P100 (95–140 ms) and VPP/N170 (145–175 ms), reflecting figure-ground segregation. After 200 ms, activity was object-sensitive during the frontocentral N3 complex (200–500 ms) implicated in cognitive decisions about visual objects, as well as a right occipitotemporal P200 (200–300 ms) implicated in perceptual grouping. All effects localized to extrastriate occipitotemporal cortex. Altogether, the findings demonstrate the time course of object-sensitive activity, which is in cortical regions defined using fMRI, and indicate that processes of figure-ground segregation (95–175 ms or longer), perceptual grouping (200–300 ms), and object model selection for a cognitive decision (200–530 ms) are achieved more for intact known objects than uncategorizable, phase scrambled versions.

Electrophysiological correlates of figure–ground segregation directly reflect perceptual saliency

In a figure identification task, we investigated the influence of different visual cue configurations (spatial frequency, orientation or a combination of both) on the human EEG. Combining psychophysics with ERP and time–frequency analysis, we show that the neural response at about 200 ms reflects perceptual saliency rather than physical cue contrast. Increasing saliency caused (i) a negative shift of the posterior P2 coinciding with a power decrease in the posterior θ-band and (ii) an amplitude and latency increase of the posterior P3. We demonstrate that visual cues interact for a percept that is non-linearly related to the physical figure–ground properties.

Probing Perceptual Antinomies with the Watercolor Illusion and Explaining How the Brain Resolves Them

The purpose of this work is to study how the brain solves perceptual antinomies, induced by the watercolor illusion in the color and in the figure-ground segregation domain, when they are present in different parts of the same object. The watercolor illusion shows two main effects: a long-range coloration and an object-hole effect across large enclosed areas (Pinna, 1987, 2005, 2008a, b; Pinna and Grossberg, 2005; Pinna et al., 2001). This illusion strongly enhances the unilateral belongingness of the boundaries (Rubin, 1915) determining grouping and figure-ground segregation more strongly than the well-known Gestalt principles. Due to the watercolor illusion, both the figure and the background assume new properties becoming, respectively, a bulging object and a hole both with a 3-D volumetric appearance (object-hole effect). When the coloration and the object-hole effects induced by the watercolor illusion are opposite (antinomic) within different portions of the same shape, some questions emerge: Do the antinomies split the shape in two parts (a half shape appears as an object and the other half as a hole) or are they solved through a new emergent perceptual result beyond the single effects? Is there a predominance of one component over the other that is less visible or totally invisible? What is perceptible and what is invisible? Is there a wholeness process under conditions where perceptual antinomies coexist? By imparting motion to a watercolored object that gradually should become a hole while overlapping another object placed behind, is the wholeness of the watercolor object weakened or reorganized in a new way? The results of phenomenological experiments suggested that the antinomies tend to be solved through two complement processes of phenomenal wholeness and partialness. These processes are explained in the light of the FACADE neural model of 3-D vision and figure-ground separation (Grossberg, 1994, 2003), notably of how complementary cortical boundary and surface representations interact with spatial attention to generate conscious percepts of 3-D form and motion.

A vision architecture for unconstrained and incremental learning of multiple categories

We present an integrated vision architecture capable of incrementally learning several visual categories based on natural hand-held objects. Additionally we focus on interactive learning, which requires real-time image processing methods and a fast learning algorithm. The overall system is composed of a figure-ground segregation part, several feature extraction methods and a life-long learning approach combining incremental learning with category-specific feature selection. In contrast to most visual categorization approaches, where typically each view is assigned to a single category, we allow labeling with an arbitrary number of shape and color categories. We also impose no restrictions on the viewing angle of presented objects, relaxing the common constraint on canonical views.

The role of vertical mirror symmetry in visualshape detection

The goal of our study is a better understanding of the role of vertical mirror symmetry in perceptual grouping. With a simple psychophysical task and a set of controlled stimuli, we investigated whether vertical mirror symmetry acts as a cue in figure-ground segregation. We asked participants to indicate which of two sequentially presented Gabor arrays contained a visual shape. The shape was defined by a subset of Gabor elements positioned along the outline of an unfamiliar shape. By adding orientation noise to these Gabor elements, the shape percept became less salient. Across the different noise levels, symmetric shapes were easier to detect than asymmetric ones. This finding indicates that vertical mirror symmetry is indeed used as a cue in perceptual grouping.

A neural model of the temporal dynamics of figure–ground segregation in motion perception

How does the visual system manage to segment a visual scene into surfaces and objects and manage to attend to a target object? Based on psychological and physiological investigations, it has been proposed that the perceptual organization and segmentation of a scene is achieved by the processing at different levels of the visual cortical hierarchy. According to this, motion onset detection, motion-defined shape segregation, and target selection are accomplished by processes which bind together simple features into fragments of increasingly complex configurations at different levels in the processing hierarchy. As an alternative to this hierarchical processing hypothesis, it has been proposed that the processing stages for feature detection and segregation are reflected in different temporal episodes in the response patterns of individual neurons. Such temporal epochs have been observed in the activation pattern of neurons as low as in area V1. Here, we present a neural network model of motion detection, figure-ground segregation and attentive selection which explains these response patterns in an unifying framework. Based on known principles of functional architecture of the visual cortex, we propose that initial motion and motion boundaries are detected at different and hierarchically organized stages in the dorsal pathway. Visual shapes that are defined by boundaries, which were generated from juxtaposed opponent motions, are represented at different stages in the ventral pathway. Model areas in the different pathways interact through feedforward and modulating feedback, while mutual interactions enable the communication between motion and form representations. Selective attention is devoted to shape representations by sending modulating feedback signals from higher levels (working memory) to intermediate levels to enhance their responses. Areas in the motion and form pathway are coupled through top-down feedback with V1 cells at the bottom end of the hierarchy. We propose that the different temporal episodes in the response pattern of V1 cells, as recorded in recent experiments, reflect the strength of modulating feedback signals. This feedback results from the consolidated shape representations from coherent motion patterns and the attentive modulation of responses along the cortical hierarchy. The model makes testable predictions concerning the duration and delay of the temporal episodes of V1 cell responses as well as their response variations that were caused by modulating feedback signals.

The electrophysiological correlate of saliency: Evidence from a figure-detection task

Although figure-ground segregation in a natural environment usually relies on multiple cues, we experience a coherent figure without usually noticing the individual single cues. It is still unclear how various cues interact to achieve this unified percept and whether this interaction depends on task demands. Studies investigating the effect of cue combination on the human EEG are still lacking. In the present study, we combined psychophysics, ERP and time–frequency analysis to investigate the interaction of orientation and spatial frequency as visual cues in a figure detection task. The figure was embedded in a matrix of Gabor elements, and we systematically varied figure saliency by changing the underlying cue configuration. We found a strong correlation between the posterior P2 amplitude and the perceived saliency of the figure: the P2 amplitude decreased with increasing saliency. Analogously, the power of the θ-band decreased for more salient figures. At longer latencies, the posterior P3 component was modulated in amplitude and latency, possibly reflecting increased decision confidence at higher saliencies. In conclusion, when the cue composition (e.g. one or two cues) or cue strength is changed in a figure detection task, first differences in the electrophysiological response reflect the perceived saliency and not directly the underlying cue configuration.

MEG recordings from human visual cortex show simultaneous complementary changes in high- and low-gamma frequency ranges to figure-ground segregation.

MEG recordings from human visual cortex show simultaneous complementary changes in high- and low-gamma frequency ranges to figure-ground segregation.

Impaired Temporal, Not Just Spatial, Resolution in Amblyopia

In amblyopia, neuronal deficits deteriorate spatial vision including visual acuity, possibly because of a lack of use-dependent fine-tuning of afferents to the visual cortex during infancy; but temporal processing may deteriorate as well. Temporal, rather than spatial, resolution was investigated in patients with amblyopia by means of a task based on time-defined figure-ground segregation. Patients had to indicate the quadrant of the visual field where a purely time-defined square appeared. The results showed a clear decrease in temporal resolution of patients' amblyopic eyes compared with the dominant eyes in this task. The extent of this decrease in figure-ground segregation based on time of motion onset only loosely correlated with the decrease in spatial resolution and spanned a smaller range than did the spatial loss. Control experiments with artificially induced blur in normal observers confirmed that the decrease in temporal resolution was not simply due to the acuity loss. Amblyopia not only decreases spatial resolution, but also temporal factors such as time-based figure-ground segregation, even at high stimulus contrasts. This finding suggests that the realm of neuronal processes that may be disturbed in amblyopia is larger than originally thought.

How Attention Affects Border Ownership in Early Visual Areas

The classic problem of figure-ground segregation draws from a more general problem of determining the side to which the border belongs. The one-sided assignment of the border to regions is termed as “border ownership” and the region to which the border is assigned is perceived as the figure ([

Figure-ground segmentation using factor graphs

Foreground–background segmentation has recently been applied [S.X. Yu, J. Shi, Object-specific figure-ground segregation, Computer Vision and Pattern Recognition (CVPR), 2003; S. Kumar, M. Hebert, Man-made structure detection in natural images using a causal multiscale random field, Computer Vision and Pattern Recognition (CVPR), 2003] to the detection and segmentation of specific objects or structures of interest from the background as an efficient alternative to techniques such as deformable templates [A.L. Yuille, Deformable templates for face recognition. Journal of Cognitive Neuroscience 3 (1) (1991)]. We introduce a graphical model (i.e. Markov random field)-based formulation of structure-specific figure-ground segmentation based on simple geometric features extracted from an image, such as local configurations of linear features, that are characteristic of the desired figure structure. Our formulation is novel in that it is based on factor graphs, which are graphical models that encode interactions among arbitrary numbers of random variables. The ability of factor graphs to express interactions higher than pairwise order (the highest order encountered in most graphical models used in computer vision) is useful for modeling a variety of pattern recognition problems. In particular, we show how this property makes factor graphs a natural framework for performing grouping and segmentation, and demonstrate that the factor graph framework emerges naturally from a simple maximum entropy model of figure-ground segmentation. We cast our approach in a learning framework, in which the contributions of multiple grouping cues are learned from training data, and apply our framework to the problem of finding printed text in natural scenes. Experimental results are described, including a performance analysis that demonstrates the feasibility of the approach.

Visual Perception and Consciousness in Dermatopathology: Mechanisms of Figure-Ground Segregation Account for Errors in Diagnosis

Visual perception has been the object of research in psychology for almost a century. Little has been written, however, about the effects of perceptive phenomena on methods in medicine that utilize interpretation of two-dimensional images for diagnosis. Starting from the work by Edgar Rubin in the beginning of the last century, this article gives a summary of observations of psychologists who investigated the mechanisms of so-called "figure-ground segregation." These unconscious mechanisms follow rules that explain why certain structures are perceived consciously as a figure, whereas other structures surrounding such a figure are neglected and not perceived consciously in detail. Perception of a structure as a figure can be due to, for example, a convex shape of its contour, proximity of lines around it, closed contours, a simple shape, and attribution of meaning to a structure. In examples from the practice of dermatopathology, those unconscious mechanisms of figure-ground segregation will be shown to be relevant to diagnosis of sections of tissue. The mechanisms help to explain why, for example, ill-defined and concave-shaped structures, stromal differences of neoplasms, interstitial infiltrates and deposits, and simulators of common diseases are often difficult to recognize at first sight. Teachers of dermatopathology need to be aware of these unconscious mechanisms of visual perception because they explain why novices struggle with certain diagnoses and differential diagnoses. Proper instruction about these phenomena, early in the process of training, will prevent a student from being frustrated with misperceptions.

Online figure–ground segmentation with adaptive metrics in generalized LVQ

We address the problem of fast figure–ground segmentation of single objects from cluttered backgrounds to improve object learning and recognition. For the segmentation, we use an initial foreground hypothesis to train a classifier for figure and ground on topographically ordered feature maps with generalized learning vector quantization. We investigate the contribution of several adaptive metrics to enable generalization to the main object parts and derive a foreground classification, which yields an improved bottom-up hypothesis. We show that metrics adaptation is a powerful enrichment, where generalizing the Euclidean metrics towards local matrices of relevance factors leads to a higher classification accuracy and considerable robustness on partially inconsistent supervised information. Additionally, we verify our results in an online learning scenario and show that figure–ground segregation using this adaptive metrics enables a considerably higher recognition performance on segmented object views.