Occluded Shapes Research Articles

Perceptual organization and visual awareness: the case of amodal completion.

We investigated the involvement of visual awareness in amodal completion, and specifically, whether visual awareness plays a differential role in local versus global completion, using a primed shape discrimination paradigm and the color-opponent flicker technique to render the prime invisible. In four experiments, participants discriminated the shape of a target preceded by a partly occluded or a neutral prime. All primes were divergent occlusion patterns in which the local completion is based on good continuation of the contours at the point of occlusion and the global completion is based on maximum symmetry. The target corresponded to the shape that could arise as a result of local or global completion of the occluded prime. For each experiment with an invisible prime we conducted a version with a visible prime. Our results suggest that local completion, but not global completion, of a partly occluded shape can take place in the absence of visual awareness, but apparently only when the visible occluded shape generates a single, local completion. No completion, either local or global, appears to take place in the absence of visual awareness when the visible occluded shape generates multiple completions. The implications of these results to the differential role of visual awareness in local and global completions and to the relationship between multiple completions and unconscious amodal completions are discussed.

Study Replication: Shape Discrimination in a Conditioning Procedure on the Jumping Spider Phidippus regius.

Spiders possess a unique visual system, split into eight different eyes and divided into two fully independent visual pathways. This peculiar organization begs the question of how visual information is processed, and whether the classically recognized Gestalt rules of perception hold true. In a previous experiment, we tested the ability of jumping spiders to associate a geometrical shape with a reward (sucrose solution), and then to generalize the learned association to a partially occluded version of the shape. The occluded shape was presented together with a broken version of the same shape. The former should be perceived as a whole shape only in the case the animals, like humans, are able to amodally complete an object partly hidden by an occluder; otherwise, the two shapes would be perceived as identical. There, the spiders learned the association but failed to generalize. Here, we present a replication of the experiment, with an increased number of subjects, a DeepLabCut-based scoring procedure, and an improved statistical analysis. The results of the experiment follow closely the direction of the effects observed in the previous work but fail to rise to significance. We discuss the importance of study replication, and we especially highlight the use of automated scoring procedures to maximize objectivity in behavioral studies.

Fast Shape Recognition Method Using Feature Richness Based on the Walking Minimum Bounding Rectangle over an Occluded Remote Sensing Target

Remote sensing target recognition has always been an important topic of image analysis, which has significant practical value in computer vision. However, remote sensing targets may be largely occluded by obstacles due to the long acquisition distance, which greatly increases the difficulty of recognition. Shape, as an important feature of a remote sensing target, plays an important role in remote sensing target recognition. In this paper, an occluded shape recognition method based on the local contour strong feature richness (contour pixel richness, contour orientation richness, and contour distance richness) to the walking minimum bounding rectangle (MBR) is proposed for the occluded remote sensing target (FEW). The method first obtains the local contour feature richness by using the walking MBR; it is a simple constant vector, which greatly reduces the cost of feature matching and increases the speed of recognition. In addition, this paper introduces the new concept of strong feature richness and uses the new strategy of constraint reduction to reduce the complex structure of shape features, which also speeds up the recognition speed. Validation on a self-built remote sensing target shape dataset and three general shape datasets demonstrate the sophisticated performance of the proposed method. FEW in this paper has both higher recognition accuracy and extremely fast recognition speed (less than 1 ms), which lays a more powerful theoretical support for the recognition of occluded remote sensing targets.

Electrophysiological evidence of the amodal representation of symmetry in extrastriate areas

Extrastriate visual areas are strongly activated by image symmetry. Less is known about symmetry representation at object-level rather than image-level. Here we investigated electrophysiological responses to symmetry, generated by amodal completion of partially-occluded polygon shapes. We used a similar paradigm in four experiments (N = 112). A fully-visible abstract shape (either symmetric or asymmetric) was presented for 250 ms (t0). A large rectangle covered it entirely for 250 ms (t1) and then moved to one side to reveal one half of the shape hidden behind (t2, 1000 ms). Note that at t2 no symmetry could be extracted from retinal image information. In half of the trials the shape was the same as previously presented, in the other trials it was replaced by a novel shape. Participants matched shapes similarity (Exp. 1 and Exp. 2), or their colour (Exp. 3) or the orientation of a triangle superimposed to the shapes (Exp. 4). The fully-visible shapes (t0–t1) elicited automatic symmetry-specific ERP responses in all experiments. Importantly, there was an exposure-dependent symmetry-response to the occluded shapes that were recognised as previously seen (t2). Exp. 2 and Exp.4 confirmed this second ERP (t2) did not reflect a reinforcement of a residual carry-over response from t0. We conclude that the extrastriate symmetry-network can achieve amodal representation of symmetry from occluded objects that have been previously experienced as wholes.

Encoding of Partially Occluded and Occluding Objects in Primate Inferior Temporal Cortex.

Object segmentation-the process of parsing visual scenes-is essential for object recognition and scene understanding. We investigated how responses of neurons in macaque inferior temporal (IT) cortex contribute to object segmentation under partial occlusion. Specifically, we asked whether IT responses to occluding and occluded objects are bound together as in the visual image or linearly separable reflecting their segmentation. We recorded the activity of 121 IT neurons while two male animals performed a shape discrimination task under partial occlusion. We found that for a majority (60%) of neurons, responses were enhanced by partial occlusion, but they were only weakly shape selective for the discriminanda at all levels of occlusion. Enhancement of IT responses in these neurons depended largely on the area of occlusion but only minimally on the color and shape of the occluding dots. In contrast to the above group of neurons, a sizable minority responded best to the unoccluded stimulus and showed strong selectivity for the shape of the discriminanda. In these neurons, response magnitude and shape selectivity declined with increasing levels of occlusion. Simulations revealed that the response characteristics of both classes of neurons were consistent with a model in which the responses to the occluded shape and the occluders are weighted separately and linearly combined. Overall, our results support the hypothesis that information about occluded and occluding stimuli are linearly separable and easily decodable from IT responses and that IT neurons encode a segmented representation of the visual scene.SIGNIFICANCE STATEMENT Recognizing partially occluded objects can be challenging, yet the primate visual system achieves it rapidly and effortlessly. For successful recognition in the face of occlusion, segmentation of the occluded and occluding objects is a critical first step. Using a combination of experimental data and simulations, here we demonstrate that responses of neurons in macaque IT cortex, the highest stage of the form processing pathway, reflect occluded and occluding stimuli as segmented components and are not bound together as they appear in the visual image. These results support the idea that segmentation and perception of occluded and occluding stimuli are directly mirrored in the responses of neurons in the highest form processing stages.

Exploring perceptual illusions in deep neural networks

Perceptual illusions -- discrepancies between what exists externally and what we actually see -- tell us a great deal about how the perceptual system functions. Rather than failures of perception, illusions reveal automatic computations and biases in visual processing that help make better decisions from visual information. Recently, deep convolutional neural networks (DCNNs) have been very successful in a variety of complex visual tasks, such as object recognition. This success has inspired researchers to begin comparing internal visual representations of DCNNs to those of humans, and in many respects, these representations turn out to be similar, raising the question of whether DCNNs “experience” some of the same illusions that people do. To investigate this, I presented a DCNN trained for object classification (VGG16, trained on imagenet) with several standard illusions, including the Muller-Lyer illusion and examples of amodal completion. Instead of using object classification, I assessed how the DCNN “perceived” the illusions by computing the similarity between the layer activation response to the ambiguous form of the illusion and to several alternate, disambiguated forms (for example, comparing the response to an occluded shape to the response to the full shape vs. to an incomplete, notched shape). For the Muller-Lyer illusion, in all convolutional layers, the response to lines with two outward tails was more similar to objectively longer lines, compared to shorter lines (p< 0.001), consistent with human perception. However, the response to lines with two inward heads was also more similar to longer lines (p< 0.001), inconsistent with human perception. For amodal completion, the response to occluded shapes was consistently more similar to the incomplete, notched shapes (p=0.003), even when the shapes were real objects. These results suggest that despite human-level performance of DCNNs on object recognition, these networks do not demonstrate some of the fundamental behavior of the visual system.

The Limiting Case of Amodal Completion: The Phenomenal Salience and the Role of Contrast Polarity.

In this work, we demonstrated unique and relevant visual properties imparted by contrast polarity in perceptual organization and in eliciting amodal completion, which is the vivid completion of a single continuous object of the visible parts of an occluded shape despite portions of its boundary contours not actually being seen. T-junction, good continuation, and closure are considered the main principles involved according to relevant explanations of amodal completion based on the simplicity–Prägnanz principle, Helmholtz’s likelihood, and Bayesian inference. The main interest of these approaches is to explain how the occluded object is completed, what is the amodal shape, and how contours of partially visible fragments are relatable behind an occluder. Different from these perspectives, amodal completion was considered here as a visual phenomenon and not as a process, i.e., the final outcome of perceptual processes and grouping principles. Therefore, the main question we addressed through our stimuli was “What is the role of shape formation and perceptual organization in inducing amodal completion?” To answer this question, novel stimuli, similar to limiting cases and instantiae crucis, were studied through Gestalt experimental phenomenology. The results demonstrated the domination of the contrast polarity against good continuation, T-junctions, and regularity. Moreover, the limiting conditions explored revealed a new kind of junction next to the T- and Y-junctions, respectively responsible for amodal completion and tessellation. We called them I-junctions. The results were theoretically discussed in relation to the previous approaches and in the light of the phenomenal salience imparted by contrast polarity.

Predictive Coding in Area V4: Dynamic Shape Discrimination under Partial Occlusion.

The primate visual system has an exquisite ability to discriminate partially occluded shapes. Recent electrophysiological recordings suggest that response dynamics in intermediate visual cortical area V4, shaped by feedback from prefrontal cortex (PFC), may play a key role. To probe the algorithms that may underlie these findings, we build and test a model of V4 and PFC interactions based on a hierarchical predictive coding framework. We propose that probabilistic inference occurs in two steps. Initially, V4 responses are driven solely by bottom-up sensory input and are thus strongly influenced by the level of occlusion. After a delay, V4 responses combine both feedforward input and feedback signals from the PFC; the latter reflect predictions made by PFC about the visual stimulus underlying V4 activity. We find that this model captures key features of V4 and PFC dynamics observed in experiments. Specifically, PFC responses are strongest for occluded stimuli and delayed responses in V4 are less sensitive to occlusion, supporting our hypothesis that the feedback signals from PFC underlie robust discrimination of occluded shapes. Thus, our study proposes that area V4 and PFC participate in hierarchical inference, with feedback signals encoding top-down predictions about occluded shapes.

Dynamic representation of partially occluded objects in primate prefrontal and visual cortex.

Successful recognition of partially occluded objects is presumed to involve dynamic interactions between brain areas responsible for vision and cognition, but neurophysiological evidence for the involvement of feedback signals is lacking. Here, we demonstrate that neurons in the ventrolateral prefrontal cortex (vlPFC) of monkeys performing a shape discrimination task respond more strongly to occluded than unoccluded stimuli. In contrast, neurons in visual area V4 respond more strongly to unoccluded stimuli. Analyses of V4 response dynamics reveal that many neurons exhibit two transient response peaks, the second of which emerges after vlPFC response onset and displays stronger selectivity for occluded shapes. We replicate these findings using a model of V4/vlPFC interactions in which occlusion-sensitive vlPFC neurons feed back to shape-selective V4 neurons, thereby enhancing V4 responses and selectivity to occluded shapes. These results reveal how signals from frontal and visual cortex could interact to facilitate object recognition under occlusion.

Object recognition under severe occlusions with a hidden Markov model approach

Shape classification has multiple applications. In real scenes, shapes may contain severe occlusions, hardening the identification of objects. In this paper, a method for object recognition under severe occlusion is proposed. Occlusion is dealt with separating shapes into parts through high curvature points, then tangent angle signature is found for each part and continuous wavelet transform is calculated for each signature. Next, the best matching object is calculated for each part using Pearson’s correlation coefficient as similarity measure between wavelets of the part and of the most probable objects in the database. For each probable class, a hidden Markov model is created in an ensemble through training with the one-class approach. For hypotheses validation, two area restrictions are set to enhance recognition performance. Experiments were conducted with 1500 test shape instances with different scenarios of object occlusion. Results showed the method not only was capable of identifying highly occluded shapes (60%–80% overlapping) but also present several advantages over previous methods.

A Mayer–Vietoris Formula for Persistent Homology with an Application to Shape Recognition in the Presence of Occlusions

In algebraic topology it is well known that, using the Mayer–Vietoris sequence, the homology of a space X can be studied by splitting X into subspaces A and B and computing the homology of A, B, and A∩B. A natural question is: To what extent does persistent homology benefit from a similar property? In this paper we show that persistent homology has a Mayer–Vietoris sequence that is generally not exact but only of order 2. However, we obtain a Mayer–Vietoris formula involving the ranks of the persistent homology groups of X, A, B, and A∩B plus three extra terms. This implies that persistent homological features of A and B can be found either as persistent homological features of X or of A∩B. As an application of this result, we show that persistence diagrams are able to recognize an occluded shape by showing a common subset of points.

Prior experience affects amodal completion in pigeons

In a three-alternative forced-choice task, 4 pigeons were trained to discriminate a target stimulus consisting of two colored shapes, one of which partially occluded the other, from two foil stimuli that portrayed either a complete or an incomplete version of the occluded shape. The dependent measure was the percentage of total errors that the birds committed to the complete foil. At the outset of training, the pigeons committed approximately 50% of total errors to the complete foil, but as training progressed, the percentage of errors to the complete foil rose. When the pigeons were given a second exposure to the initial set of stimuli, they committed 70% of total errors to the complete foil, suggesting that they now saw the complete foil as more similar to the occluded target than the incomplete foil. These results suggest that experience with 2-D images may facilitate amodal completion in pigeons, perhaps via perceptual learning.

Better discrimination for illusory than for occluded perceptual completions

We applied the thin-fat Kanizsa shape discrimination task invented by D. L. Ringach and R. Shapley (1996) to study perceptual completion by measuring whether the discrimination was more accurate for illusory than for occluded shapes. Differently from Ringach and Shapley, we tested naive observers with stereoscopic displays. Discrimination was consistently more accurate for illusory than for occluded shapes under a variety of stimulus conditions. However, the absolute performance was worse than Ringach and Shapley's experienced observers, who discriminated illusory and occluded shapes equally well. When our naive observers were trained, their performance approached that in Ringach and Shapley, and their performance difference diminished between the illusory and occluded. The more precise discrimination of the illusory shapes by untrained observers is consistent with the subjective impression that illusory contours appear clearer and positionally better defined. This makes sense from the perspective of Bayesian decision theory: the location of an illusory contour that is closer to an observer might be more important than an occluded contour, and hence obligatorily represented more precisely. We conclude the paper by discussing implications of our results on the current debate on mechanisms of perceptual completion (M. K. Albert, 2007; B. L. Anderson, 2007; P. J. Kellman, P. Garrigan, T. F. Shipley, & B. P. Keane, 2007).

A fast evaluation criterion for the recognition of occluded shapes

A robust and fast object recognition method is crucial in many robotic applications, especially (but not restricted to) in manufacturing. This paper introduces a novel algorithm that satisfies both of these criteria and is capable of recognising a set of model shapes in a complicated scene given as input. The shapes are described using the turning angle representation. Shape matching is carried out by finding the correspondence between the model shape and the input (i.e. corresponding points), and calculating the geometrical transformation of the model that minimises the least square distance between the corresponding points. A new geometric feature is proposed, called “Generalised Angle”, which facilitates fast elimination of infeasible matches. The Generalised Angle ( GA ) is invariant to rotation, translation and scaling and does not result in a considerable computational cost to the system. Moreover, an evaluation function is used, which takes several criteria into account and renders the method capable of recognising shapes under occlusion effectively.

The Occlusion Illusion: Partial Modal Completion or Apparent Distance?

In the occlusion illusion, the visible portion of a partly occluded object (eg a semicircle partly hidden behind a rectangle) appears to be significantly larger than a physically identical region that is fully visible. This illusion may occur either because the visual system 'fills in' a thin strip along the occluded border (the partial-modal-completion hypothesis) or because the partly occluded object is perceived as farther away (the apparent-distance hypothesis). We measured the magnitude of the occlusion illusion psychophysically in several experiments to investigate its causes. The results of experiments 1-3 are consistent with the general proposal that the magnitude of the illusion varies with the strength of the evidence for occlusion, supporting the inference that it is due to occlusion. Experiment 4 provides a critical test between apparent-distance and partial-modal-completion explanations by determining whether the increase in apparent size of the occluded region results from a change in its perceived shape (due to the modal extension of the occluded shape along the occluding edge, as predicted by the partial-modal-completion hypothesis) or from a change in its perceived overall size (as predicted by the apparent-distance hypothesis). The results more strongly support the partial-modal-completion hypothesis.

Prior experience affects amodal completion in pigeons

In a three-alternative forced-choice task, 4 pigeons were trained to discriminate a target stimulus consisting of two colored shapes, one of which partially occluded the other, from two foil stimuli that portrayed either a complete or an incomplete version of the occluded shape. The dependent measure was the percentage of total errors that the birds committed to the complete foil. At the outset of training, the pigeons committed approximately 50% of total errors to the complete foil, but as training progressed, the percentage of errors to the complete foil rose. When the pigeons were given a second exposure to the initial set of stimuli, they committed 70% of total errors to the complete foil, suggesting that they now saw the complete foil as more similar to the occluded target than the incomplete foil. These results suggest that experience with 2-D images may facilitate amodal completion in pigeons, perhaps via perceptual learning.

Investigating metrical and structural aspects of visual completion: Priming versus searching

Earlier research on visual occlusion showed some flexibility in the formation of visual completions, as long as the structural aspects (e.g., symmetry) of the visible part of the partly occluded shape were preserved in the completion (de Wit & van Lier, 2002). In this study, we examined whether changing the overall size of the occluded shape would preserve the overall structure. In Experiment 1, using the primed‐matching paradigm, we found evidence for relative size invariance in the completion process. To investigate whether changes in the structural aspects of shape are generally more salient than those of size, we employed the same stimuli in visual search and change detection paradigms. Experiment 2 demonstrated effects of completion in both paradigms. Experiment 3 showed that the metrical aspects of the shapes used in Experiment 1 are nevertheless detected faster than the structural aspects under search conditions. Because the variation in structural aspects is not more salient than in metrical aspects, we conclude that for these shapes, visual completion is indeed size‐invariant. The relations between performances in the three paradigms are discussed.

Fast Retrieval of Isolated Visual Shapes

Similarity-based retrieval from databases of isolated visual shapes has become an important information retrieval problem. The goal of the current work is to achieve high retrieval speed with reasonable retrieval effectiveness, and support for partial and occluded shape queries. In the proposed method, histograms of local shape parts are coded as index vectors. To increase retrieval accuracy, a rich set of parts at all scales of the shape is used; specifically, the parts are defined as connected sequences of regions in curvature scale space. To increase efficiency, structural indexing is used to compare the index vectors of the query and database shapes. In experimental evaluations, the method retrieved at least one similar shape in the top three retrieved items 99–100% of the time, depending on the database. Average retrieval times ranged from 0.7 ms on a 131-shape database to 7 ms on a 1310-shape database. The method is thus suitable for fast, approximate shape retrieval in comparison with more accurate but more costly structural matching.

Visuo-cognitive disambiguation of occluded shapes

Pessoa et al. (1998a) underexposed the broad and rich variety of stimuli in the amodal completion domain. The disambiguation of occluded shapes depends on very specific figural properties. Elaborations on such disambiguations of rich and complex stimuli, tied up with a visuo-cognitive origin of amodal completion, further position Pessoa et al.'s considerations on neural filling-in and the personal-subpersonal distinction.

The computation of occluded contours in visual agnosia: Evidence for early computation prior to shape binding and figure-ground coding

We examined whether an agnosic patient with a deficit in early visual processing, HJA, completed occluded contours. We used matching tasks with stimuli composed of three superimposed or occluded shapes. Experiments 2 and 6 required superimposed or occluded shapes to be discriminated from distractors in which the position of one shape was changed. HJA was selectively impaired with occluded relative to superimposed shapes. His performance was affected by the spatial separation of the occluded contours rather than the area of the occluded surface. Experiments 3 and 5 required HJA to discriminate the central shape. Making occluded contours easier to compute (by reducing their spatial separation) facilitated discrimination of a central occluded shape (in the background), although it impaired discrimination of a central occluding shape (in the foreground). Free-choice shape judgements made to the central shape (Experiment 2) showed that HJA used both real and completed contours to segment foreground shapes inappropriately. When asked to copy overlapping shapes (Experiment 4), HJA drew in the occluded parts as if real contours were present, at least on some occasions. These drawings and a task requiring discrimination between real and occluded contours (Experiment 7), showed a tendency to continue contours inappropriately, an insensitiviy to junctions, and impaired integration of contours into more global shapes. The results suggest that occluded contours can be computed early on in visual processing, probably at the level where long-range mechanisms group collinear contour segments together. Our control experiment shows that HJA is not impaired in collinear contour grouping. These mechanisms are prior to processes in which contours are bound to shapes and in which foregroundbackground relationships between shapes are resolved. In visual agnosia, occluded contours can be computed even when there is impairment of both binding of contours to shapes and the computation of foreground-background relations in overlapping shapes.