Subjective Contours Research Articles

随机修复场模型的改进<br>Improvement on the Stochastic Completion Field

主观轮廓是一种重要的视觉心理现象，体现了人类视觉惊人的感知修复能力，而随机修复场模型则是针对主观轮廓提出的一个比较有影响力的计算模型。针对随机修复场模型存在的问题提出了一个改进方案。该方案建立了一个根据图像边缘的位置和取向构成的几何关系提取主观轮廓的综合性解决方法。实验结果验证了算法的有效性。 Recognizing subjective contour is an important visual psychological phenomena, which demonstrating remarkable capability of human vision. Among the related works on subjective contour extraction, the stochastic completion fields model stands out. However, there are still some computation problems of this model which are still unsolved. This paper presents a perceptual completion method for subjective contour to overcome these problems. The algorithms suggested use edge evidences and local orientations to constrain possible geometrical groupings and completions and experimental results show its superiorities.

Poggendorff illusion with subjective contours

Poggendorff illusion with subjective contours

Priming effect on the subjective contour of a cat's figure

Priming effect on the subjective contour of a cat's figure

Organização espacial na percepção visual de luminosidade

Este estudo analisou a influência de variações físicas dos estímulos, com base na organização espacial de figura-fundo criada pela associação dos efeitos ilusórios de contraste simultâneo de luminosidade e de contornos subjetivos. A cada participante, no total de 64, foram apresentadas 160 matrizes de escolha, cada uma composta de um estímulo modelo e quatro estímulos de comparação, devendo ser identificado qual dos quatro estímulos de comparação correspondia ao estímulo modelo. A diferença significativa entre as médias de ajuste visual verificadas para a condição de contorno subjetivo médio e para a condição controle (sem contorno) mostrou que a formação clássica de contornos subjetivos de Kanizsa, quando associada ao efeito de contraste simultâneo de luminosidade, influenciou a percepção de luminosidade dos participantes.

Visual constraints for the perception of quantitative depth from temporal interocular unmatched features

Previous research ( Brooks & Gillam, 2006) has found that temporal interocular unmatched (IOUM) features generate a perception of subjective contours and can result in a perception of quantitative depth. In the present study we examine in detail the factors important for quantitative depth perception from IOUM features. In Experiments 1 and 2 observers were shown temporal IOUM features based on three dots that disappeared behind an implicit surface. Subjects reported a perception of a subjective surface and were able to perceive qualitative depth. In Experiments 3 and 4 metrical depth was perceived when binocular disparity features were added to the display. These results suggest that quantitative depth from IOUM information is perceived when binocular matched information is present in regions adjacent to the surface. In addition, the perceived depth of the subjective surface decreased with an increase in the width of the subjective surface suggesting a limitation in the propagation of quantitative depth to surface regions where qualitative depth information is available.

Subjective contours and binocular rivalry

Subjective contours and binocular rivalry

Towards a mathematical theory of cortical micro-circuits.

The theoretical setting of hierarchical Bayesian inference is gaining acceptance as a framework for understanding cortical computation. In this paper, we describe how Bayesian belief propagation in a spatio-temporal hierarchical model, called Hierarchical Temporal Memory (HTM), can lead to a mathematical model for cortical circuits. An HTM node is abstracted using a coincidence detector and a mixture of Markov chains. Bayesian belief propagation equations for such an HTM node define a set of functional constraints for a neuronal implementation. Anatomical data provide a contrasting set of organizational constraints. The combination of these two constraints suggests a theoretically derived interpretation for many anatomical and physiological features and predicts several others. We describe the pattern recognition capabilities of HTM networks and demonstrate the application of the derived circuits for modeling the subjective contour effect. We also discuss how the theory and the circuit can be extended to explain cortical features that are not explained by the current model and describe testable predictions that can be derived from the model.

Details of the construction of perception: a closer look at illusory contours

Are we aware of neural activity in primary visual cortex? This question was asked in a classic paper written by Nobel laureate Francis Crick and Christof Koch more than a decade ago (Crick and Koch, 1995). By now, several studies have addressed the question, but so far consensus has not been reached (Rees, 2007; Tong, 2003). A recent study by Maertens et al. (2008) presents new relevant data on this issue. They demonstrated that activity in primary visual cortex (V1) is correlated with the perception of subjective contours. Maertens et al. used Kanizsa figures to induce a percept of illusory contours (IC, Figure ​Figure1),1), which are subjective in the sense that there is no actual change in luminance at the location of the perceived contour. Figure 1 Kanizsa stimuli used by Maertens et al. (2008). Illusory contours were generated by inducers (black “pack-men”) shown at different angles. The task was to judge whether the illusory contour was convex or concave. As control condition, ... The (illusory) perception of the contours is dependent on coherent orientation of the inducers. Hence, by either aligning or misaligning the inducers, very similar (local) physical parameters can create two subjectively different visual experiences. IC can therefore be used to provide information on neural correlates of conscious perception. An important feature of Maertens et al.’s experiment was the localization of neural activity specific for contour perception. This was accomplished by using spatially specific localizer stimuli for the contours and inducers separately, as well as by using a spatially more precise imaging sequence than traditionally used ones. This enabled the conclusion that the IC were related to retinotopically specific regions within V1, thereby extending previous research on IC where IC-related activity at such a level of detail was not observed. The results also add to existing evidence regarding V1 involvement in conscious perception. In previous research on conscious perception, some but not other studies have observed correlated activity in V1. A factor that could account for the inconsistent results is the specific task requirements. Specifically, it has been suggested that there is a reverse hierarchical sequence for conscious perception such that global, holistic features are perceived first and the details are added if required (Hochstein and Ahissar, 2002). In the Maertens et al. paper, the angle of the induced contrasts was varied between trials and the participant was required to judge the angle on each trial (Figure ​(Figure1).1). Hence, the task mandated a fine-scaled decision which required detail to be extracted – detail at such a fine scale that it may only be present in V1 due to its small receptive fields – and might therefore have induced correlated V1 activity. A hypothesis put forth by Crick and Koch in their 1995 paper was that, while activity in V1 may be necessary for conscious perception, it may not be sufficient. According to Crick and Koch, sensory regions need to be connected to prefrontal cortex to enable conscious perception. Whether involvement of frontal regions is necessary remains controversial (Eriksson et al., 2008), but a role of prefrontal cortex is indicated by a large number of neuroimaging studies showing a correlation between activity in frontal regions and conscious perception (Naghavi and Nyberg, 2005; Rees et al., 2002). In the Maertens et al. study, a limited part of the brain was sampled to improve spatial resolution. While this enabled the retinotopic mapping of the IC and thus constituted a great strength of study, this level of detail comes with a cost as nothing could be concluded regarding frontal involvement in the creation of IC. Future work, building on the methods used in the Maertens et al. study along with whole-brain sampling could shed light on the involvement of higher-order cortical regions and necessary vs. sufficient conditions for conscious perception. In sum, the Maertens et al. study is an excellent demonstration of the constructive nature of perception (von Helmholtz, 1910). Moreover, it shows that this constructive process is not confined to higher levels of the visual hierarchy, but can take place even at the lowest level of cortical processing.

The perception of subjective contours and neon color spreading figures in young infants

The goal of the present habituation-dishabituation study was to explore sensitivity to subjective contours and neon color spreading patterns in infants. The first experiment was a replication of earlier investigations that showed evidence that even young infants are capable of perceiving subjective contours. Participants 4 months of age were habituated to a subjective Kanizsa square and were tested afterward for their ability to differentiate between the subjective square and a nonsubjective pattern that was constructed by rotating some of the inducing elements. Data analysis indicated a significant preference for the nonsubjective pattern. A control condition ensured that this result was not generated by the difference in figural symmetry or by the local differences between the test displays. In the second experiment, infant perception of a neon color spreading display was analyzed. Again, 4-month-old infants could discriminate between the illusory figure and a nonillusory pattern. Furthermore, infants in a control group did not respond to the difference in symmetry and the local differences between two nonillusory targets. Overall, the results show that young infants respond to illusory figures that are generated by either implicit T-junctions (Experiment 1) or implicit X-junctions (Experiment 2). The findings are interpreted against the background of the neurophysiological model proposed by Grossberg and Mingolla (1985).

Perception of Subjective Contours in Fish

The ability of fish to perceive subjective (or illusory) contours, ie contours that lack a physical counterpart in terms of luminance contrast gradients, was investigated. In the first experiment, redtail splitfins (Xenotoca eiseni), family Goodeidae, were trained to discriminate between a geometric figure (a triangle or a square) on various backgrounds and a background without any figure. Thereafter, the fish performed test trials in which illusory squares or triangles were obtained by (i) interruptions of a background of diagonal lines, (ii) phase-shifting of a background of diagonal lines, and (iii) pacmen spatially arranged to induce perception of Kanizsa subjective surfaces. In all three conditions, fish seemed to generalise their responses to stimuli perceived as subjective contours by humans. Fish chose, correctly, squares or triangles made of interrupted or phase-shifted diagonal lines from uniform backgrounds of diagonal lines, as well as illusory square or triangle Kanizsa figures from figures in which the inducing pacmen were scrambled. In the second experiment, fish were trained to discriminate between a vertical and a horizontal bar with luminance contrast gradients, and then tested with vertically and horizontally oriented illusory bars, created either through interruption or spatial phase-shift of inducing diagonal lines. Fish appeared to be able to generalise the orientation discrimination to illusory contours. These results demonstrate that redtail splitfins are capable of perceiving illusory contours.

Nakayama, Shimojo, and Ramachandran's 1990 Paper

The target paper reviewed in this article was titled Transparency: relation to depth, subjective contours, luminance, and neon colour spreading coauthored by K Nakayama, S Shimojo, and V S Ramachandran, published in 1990. This paper, one of the first in a series on surface perception, examined how in untextured stereograms, local disparity and luminance contrast can drastically change surface quality, subjective contours, and the effect of neon colour spreading. When we began to conceive this and related work, the ascendant view on visual perception was derived from the pioneering studies of the response properties of visual neurons with microelectrodes, including those of Barlow (1953), Lettvin et al (1959), and Hubel and Wiesel (1959, 1962). All suggested that there are remarkable operations on the image by earliest stages of the visual pathway, which bestowed selectivity to colour, orientation, motion direction, spatial frequency, binocular disparity, etc. As such, it would seem that an understanding of vision would come through more systematic description of the properties of single neuron selectivities. This viewpoint was well summarised by Horace Barlow in his famous neuron doctrine paper (Barlow 1972), which emphasised the importance of analysing the image in successive stages of processing by neurons with specific classes of receptive fields. Later work altered this conception somewhat by seeing receptive fields as linear filters. Rather than detecting the presence of edges, bars, or otherwise perceptually identifiable elements in a scene, cells were seen as making measurements of an image. It was an 'image based' approach to vision, treating the basic operations of vision with no particular regard as to what aspects of scenes were being coded, whether a given cell's response corresponded to something about surfaces, edges, or objects in the real world.

Poggendorff illusion with Kanizsa-like subjective contours

Poggendorff illusion with Kanizsa-like subjective contours

Surround Suppression of V1 Neurons Mediates Orientation-Based Representation of High-Order Visual Features

Neurons with surround suppression have been implicated in processing high-order visual features such as contrast- or texture-defined boundaries and subjective contours. However, little is known regarding how these neurons encode high-order visual information in a systematic manner as a population. To address this issue, we have measured detailed spatial structures of classical center and suppressive surround regions of receptive fields of primary visual cortex (V1) neurons and examined how a population of such neurons allow encoding of various high-order features and shapes in visual scenes. Using a novel method to reconstruct structures, we found that the center and surround regions are often both elongated parallel to each other, reminiscent of on and off subregions of simple cells without surround suppression. These structures allow V1 neurons to extract high-order contours of various orientations and spatial frequencies, with a variety of optimal values across neurons. The results show that a wide range of orientations and widths of the high-order features are systematically represented by the population of V1 neurons with surround suppression.

Local determinants of contour interpolation

Objects in our visual environment are perceived as integral wholes even when their retinal images are incomplete. We ask whether the perceptual precision of subjective interpolation between isolated image parts depends on the overall proportion of visible image information or rather on its geometrical arrangement. We used Varin-type subjective shapes that provide less physical stimulus information than Kanizsa-type figures because partially occluded solid inducers are replaced by partially occluded concentric arcs. We tested whether perceptual precision varies as a function of contour support, or alternatively, depends on the number of, and the distance between, line endings within the inducers. We measured performance in a probe localization task, where a small target is presented at different distances around a subjective boundary. Sensitivity, captured by the just noticeable position difference between in- and outside probes, crucially depended on the geometric arrangement of line ends in the Varin figures. This is objective evidence that the apparent subjective contour strength does not primarily depend on contour support but is determined by the number and the separation between inducers' line endings. The results suggest that neuronal mechanisms sensitive to highly localized 2D features are crucial for determining the perceived shape of visual objects.

Kanizsa-type subjective contours do not guide attentional deployment in visual search but line termination contours do

We used visual search to explore whether attention could be guided by Kanizsa-type subjective contours and by subjective contours induced by line ends. Unlike in previous experiments, we compared search performance with subjective contours against performance with real, luminance contours, and we had observers search for orientations or shapes produced by subjective contours, rather than searching for the presence of the contours themselves. Visual search for one orientation or shape among distractors of another orientation or shape was efficient when the items were defined by luminance contours. Search was much less efficient among items defined by Kanizsa-type subjective contours. Search remained efficient when the items were defined by subjective contours induced by line ends. The difference between Kanizsa-type subjective contour and subjective contours induced by line ends is consistent with physiological evidence suggesting that the brain mechanisms underlying the perception of these two kinds of subjective contours may be different.

Biases and Sensitivities in the Poggendorff Effect when Driven by Subjective Contours

A consensus in the existing literature suggests that the Poggendorff effect (a perceptual misalignment of two collinear transversal segments when separated by a pair of parallel contours) persists when the parallels are defined by Kanizsa-like subjective contours. However, previous studies have often been complicated by a lack of quantitative measures of effect size, statistical tests of significance, appropriate measures of baseline and control biases, or stringent definition of subjective contours. The aim of this study was thus to determine whether subjective contours are capable of driving the Poggendorff effect once other factors are accounted for. Twenty participants were tested on a number of test and control figures incorporating first-order (luminance-defined) and subjective parallels using the method of adjustment. All figures were tested at two different orientations, and observer sensitivities and observer biases were assessed. A systematic response bias (in the direction of the classical effect) was found for Poggendorff figures that incorporated subjective parallels. The effect was highly significant and greater than for control figures. There was no concomitant change in judgment sensitivity (positional certainty). Finally, there was a positive correlation between the effect size for figures incorporating first-order and subjective parallels. The findings reported demonstrate conclusively that true Kanizsa-like subjective contours are capable of driving the Poggendorff effect. Further, the data are consistent with a growing body of evidence that suggests both first-order and subjective contours are processed at early loci in the visual pathways when position is encoded.

Retinotopic activation in response to subjective contours in primary visual cortex

Objects in our visual environment are arranged in depth and hence there is a considerable amount of overlap and occlusion in the image they generate on the retina. In order to properly segment the image into figure and background, boundary interpolation is required even across large distances. Here we study the cortical mechanisms involved in collinear contour interpolation using fMRI. Human observers were asked to discriminate the curvature of interpolated boundaries in Kanizsa figures and in control configurations, which contained identical physical information but did not generated subjective shapes. We measured a spatially precise spin-echo BOLD signal and found stronger responses to subjective shapes than non-shapes at the subjective boundary locations, but not at the inducer locations. The responses to subjective contours within primary visual cortex were retinotopically specific and analogous to that to real contours, which is intriguing given that subjective and luminance-defined contours are physically fundamentally different. We suggest that in the absence of retinal stimulation, the observed activation changes in primary visual cortex are driven by intracortical interactions and feedback, which are revealed in the absence of a physical stimulus.

Infants’ perception of subjective contours from apparent motion

We examined infants’ perception of subjective contours in Subjective-Contour-from-Apparent-Motion (SCAM) stimuli [e.g., Cicerone, C. M., Hoffman, D. D., Gowdy, P. D., & Kim, J. S. (1995). The perception of color from motion. Perception & Psychophysics, 57, 761–777] using the preferential looking technique. The SCAM stimulus is composed of random dots which are assigned two different colors. Circular region assigned one color moved apparently, keeping all dots’ location unchanged. In the SCAM stimulus, adults can perceive subjective color spreading and subjective contours in apparent motion ( http://c-faculty.chuo-u.ac.jp/∼ymasa/okamura/ibd_demo.html). In the present study, we conducted two experiments by using this type of SCAM stimulus. A total of thirty-six 3–8-month-olds participated. In experiment 1, we presented two stimuli to the infants side by side: a SCAM stimulus consisting of different luminance, and a non-SCAM stimulus consisting of isoluminance dots. The results indicated that the 5–8-month-olds showed preference for the SCAM stimuli. In experiments 2 and 3, we confirmed that the infants’ preference for the SCAM stimulus was not generated by the local difference and local change made by luminance of dots but by the subjective contours. These results suggest that 5–8-month-olds were able to perceive subjective contours in the SCAM stimuli.

수정된 캐니 에지 맵으로부터 만들어진 LOD 에지 맵을 이용한 물체 추적 및 소거

본 논문은 하나의 움직이는 카메라와 수시로 바뀌는 배경을 가진 환경 하에서 파라미터를 사용하지 않는 외곽선을 사용한 움직이는 물체의 외곽을 추적하고, 추적된 물체의 외곽을 다른 장면에서 가져온 배경으로 대체하여 추적물체를 제거하는 기법을 제안한다. 먼저 캐니 에지 이미지(map)를 수정하여 만들어 내고, 이들 에지들의 강도에 대한 정보를 LOD (Level-of-Detail)로 만든 결과 LOD 캐니 에지 이미지(map)을 생성한다. 이들 LOD 캐니 에지 이미지 화소에 대해 그래프를 사용한 경로 설정 방법을 사용한다. 이 작업으로 결정되는 외곽선을 이용하여 추적대상이 되는 물체를 다른 이미지에서부터 얻은 배경이미지로 대체함으로써 제거한다. 우리의 물체 추적을 위한 방법은 LOD 수정된 캐니에지 이미지를 위주로 이루어진다. 추가 에지 정보를 얻기 위해 LOD 계층에 따라서 자세한 외곽선 정보를 얻는다. 우리의 경로 설정 방법은 보다 강한 이미지 차에서 만들어진 에지 화소를 선호하는 것이다. 이 방법은 이전 외곽선 정보를 최소한으로 참고하기 때문에, 이전 외곽선 정보를 새로운 외곽선을 생성하는데 있어서 가중치를 사용 이전 외곽선을 포함시키는 방법에 비해 탁월하다. 외곽선 추적 후 추적 물체를 배경으로 대체하는데, 첫 이미지 배경은 이후에 나타나는 이미지로부터 추적 물체에 대해 가려진 배경정보를 가져오는 카메라 운동법이라 부르는 방법에 의하여 계산되어진다. 첫 프레임을 위한 배경 계산이 완료되면, 다음 이미지의 배경 계산은 첫 프레임의 배경에 의존한다. 본 논문에서 제시된 방법을 사용할 경우, 추적 물체의 형상 변화가 극심하지 않고, 카메라의 움직임이 매우 빠르지 않을 경우 성공적으로 추적할 수 있었다. We propose a simple method for tracking a nonparameterized subject contour in a single video stream with a moving camera and changing background. Then we present a method to eliminate the tracked contour object by replacing with the background scene we get from other frame. First we track the object using LOD (Level-of-Detail) canny edge maps, then we generate background of each image frame and replace the tracked object in a scene by a background image from other frame that is not occluded by the tracked object. Our tracking method is based on level-of-detail (LOD) modified Canny edge maps and graph-based routing operations on the LOD maps. We get more edge pixels along LOD hierarchy. Our accurate tracking is based on reducing effects from irrelevant edges by selecting the stronger edge pixels, thereby relying on the current frame edge pixel as much as possible. The first frame background scene is determined by camera motion, camera movement between two image frames, and other background scenes are computed from the previous background scenes. The computed background scenes are used to eliminate the tracked object from the scene. In order to remove the tracked object, we generate approximated background for the first frame. Background images for subsequent frames are based on the first frame background or previous frame images. This approach is based on computing camera motion. Our experimental results show that our method works nice for moderate camera movement with small object shape changes.

Mislocalization of a target toward subjective contours: attentional modulation of location signals

This study examined whether a briefly presented target was mislocalized toward a subjective contour. Observers manually reproduced the position of a briefly presented peripheral target circle above a central fixation cross. A luminance contour, a subjective contour, or a no-contour stimulus was presented in either the left of right visual field, and a no-contour control was presented in the opposite visual field. After these stimuli vanished, a target circle was then presented. Consequently, the degree of mislocalization toward the subjective and luminance contours was the same; this indicated that image integration at a coarse spatial scale cannot explain mislocalization. Experiment 2 revealed that the mislocalization in Experiment 1 was not a result of eye movements. Experiment 3 found that the spatial attention allocated at the location of the luminance and subjective contours was more than that allocated at the no-contour stimulus. An attentional shift toward the task-irrelevant stimulus resulted in a mislocalization of the target.