Title Participants Abstract "50 Years of Stereoblindness: Reconciliation of a Continuum of Disparity Detectors With Blindness for Disparity in Near or Far Depth" "Raymond van Ee" "Whitman Richards (1932-2016) discovered some 50 years ago that about 30% of observers from the normal population exhibit stereoblindness: the disability to process binocular disparities in either far or near depth. We review the literature on stereoblindness entailing two insights. First, contemporary scholars in stereopsis undervalue the comprehension that disparity processing studies require precise assessments of observers' stereoblindness. We argue that this frequently leads to suboptimal interpretations. Second, there is still an open conundrum: How can the established finding that disparity is processed by a continuum of detectors be reconciled with the disability of many observers to process a whole class of far or near disparities? We propose, based upon integration of literature, that an asymmetry between far and near disparity detection at birth-being present for a variety of reasons-can suppress the typical formation of binocular correlation during the critical period for the development of stereopsis early in life, thereby disabling a whole class of far or near disparities." "The shifting focus and assumptions in disparity research" "Olga Petintseva" "This paper discusses shifts in research on disparity throughout judicial trajectories of ethnic minorities and immigrants (both disparity through prosecutorial power and judicial disparity). After a brief overall overview with attention to turns in the definition of disparity, target groups and levels of analysis, we particularly focus on the underlying ontological assumptions about agency of decision makers as individual reflexive actors versus the structuralist vision, that have been going back and forth between these two extremes of a continuum for several decades and whereby different scholars have attempted to integrate the seemingly contradictory visions. This contribution attempts to make these assumptions - particularly in European disparity research explicit, to discuss their impact and to identify the shifts of the past three decades." "Real-Time Edge-Sensitive Local Stereo Matching with Iterative Disparity Refinement" "Maarten DUMONT, Patrik GOORTS, Steven MAESEN, Gauthier LAFRUIT, Philippe BEKAERT" "First, we present a novel cost aggregation method for stereo matching that uses two edge-sensitive shape-adaptive support windows per pixel region; one following the horizontal edges in the image, the other the vertical edges. Their combination defines the final aggregation window shape that closely follows all object edges and thereby achieves increased hypothesis confidence. Second, we present a novel iterative disparity refinement process and apply it to the initially estimated disparity map. The process consists of four rigorously defined and lightweight modules that can be iterated multiple times: a disparity cross check, bitwise fast voting, invalid disparity handling, and median filtering. We demonstrate that our iterative refinement has a large effect on the overall quality, resulting in smooth disparity maps with sharp object edges, especially around occluded areas. It can be applied to any stereo matching algorithm and tends to converge to a final solution. Finally, we perform a quantitative evaluation on various Middlebury datasets, showing an increase in quality of over several dB PSNR compared with their ground truth. Our whole disparity estimation algorithm supports efficient GPU implementation to facilitate scalability and real-time performance." "fMRI guided Single-cell studies of higher-order disparity selectivity in extrastriate cortex of the macaque brain" "Amirmohammad Alizadeh" "Introduction Depth perception is a critical and demanding duty that our brain has to carry out every day. Among all depth cues, horizontal binocular disparity is one of the strongest and a non-ambiguous depth cue. The differences between the locations of matching features on the retinae are termed binocular disparities and they give rise to a vivid percept of depth. All real life objects exhibit three principal levels of disparity orders: zero order disparity reflects the object’s position in depth relative to the fixation plane (absolute disparity, near or far), whereas higher-order disparities contain changes of disparity along a surface. First-order disparity refers to linear gradients of disparity, as in tilted planar surfaces, whereas second-order disparity indicates surfaces that are curved in depth In the primate brain, both the ventral pathway (which is important for object recognition and categorization) and the dorsal visual pathway (which is important for spatial vision and actions) contain several areas that encode depth information from disparity. Both visual pathways originate in the primary visual cortex (V1), which already contains neurons that respond selectively to zero order disparity (Cumming and Parker, 1999). Higher-order disparity selectivity in the ventral stream was initially discovered in a small region in the rostral part of the lower bank of the Superior Temporal Sulcus (STS), labeled TEs by Janssen et al. (2000a) who showed that neurons can be selective for very small differences in depth structure (Janssen et al., 2000b), for first- and second order disparities (Janssen et al., 2000b) and for three-dimensional surfaces and boundaries (Janssen et al., 2001). Use of an evolutionary stimulus strategy (stimuli rendered with shading and binocular disparity) suggests that neurons in inferotemporal cortex are simultaneously tuned for 3D shape (maximum and minimum principle surface curvatures), 3D orientation, and relative 3D position of the constituent surface fragments (Yamane et al., 2008). The dorsal visual pathway is directed towards the Intraparietal Sulcus (IPS), ending in the premotor regions in frontal cortex. The caudal intraparietal area (CIP) in the lateral bank of the IPS has been studied using first-order stimuli defined by disparity and texture gradients (Shikata et al., 1996; Taira et al., 2000; Tsutsui et al., 2001; Durand et al., 2009; Rosenberg et al., 2013). Area AIP (located in the lateral bank of the IPS anterior to CIP) represents an important locus for first and second order disparity processing and the output of this area is believed to feed into the ventral premotor cortex, providing 3D object information for grasping (Murata et al., 2000; Srivastava et al., 2009; Theys et al., 2012a; Theys et al., 2012b). General hypothesis and specific aims of the project The overall goal of this project is to investigate the neural selectivity for higher-order disparities in the depth structure network in both dorsal and ventral visual stream, with a focus on the mid-level areas PIP (dorsal stream) and TEO (ventral stream) . Our general hypothesis is that the neural representation of depth structure differs between the two main visual processing streams. Our study will provide the first detailed comparison between these two areas with respect to their neuronal properties and their role in depth perception. Experiment 1. Single-cell properties in area PIP: In one of their studies, Tsutsui et al. (2002) reported the existence of 3D responses in a very small number of recording sites in the medial bank of the caudal part of the IPS (area PIP). In addition, a recent study in our laboratory (Premereur et al., 2015) showed that electrical microstimulation of 3D-shape selective clusters of neurons in posterior AIP during functional Magnetic Resonance Imaging (fMRI) activates CIP and PIP. In a separate set of studies, strong fMRI activation was observed in area PIP in response to curved stimuli compared to flat stimuli in our laboratory. Therefore, the goal of this experiment is to investigate single unit activity of PIP neurons in response to higher order disparities embedded in random dot stereograms. We want to precisely define details of higher-order responses in PIP such as selectivity for planar and curved surfaces, size tolerance, receptive field profile, etc. Currently a uniform and systematic study of the response properties of disparity selective cells in PIP is not available. Experiment 2. Single-cell properties in area TEO Janssen and colleagues studied single cell responses of the TEs region in inferotemporal cortex (Janssen et al., 1999; Janssen et al., 2000a; Janssen et al., 2001). TEs contains a very detailed representation of 3D shape, but ongoing imaging studies in our laboratory suggest that intermediate areas along the ventral stream (e.g. TEO) may also contain neurons that respond selectively to higher-order disparities. In this experiment, we will explore the neural selectivity of TEO neurons to mean disparity (zero order) as well as disparity gradients (higher order). This will include a detailed study of response properties such as size tolerance and receptive field measurements. Experiment 3. Behavioral role of areas PIP and TEO. To investigate the role of PIP and TEO in depth perception, we will apply electrical microstimulation in clusters of neurons showing higher-order disparity selectivity. We will first record single-unit and multi-unit activity over a range of recording depths, and if clustering of the selectivity is observed we will electrically stimulate in half of the trials while the animals perform a depth structure categorization task (Verhoef et al., 2012). Our working hypothesis is that we will be able to elicit perceptual effects when stimulating TEO, but not when stimulating PIP neurons. Methodology The experiments described here were approved by the ethical committee of the KU Leuven. Because disparity naturally requires frontal placement of the eyes, only a primate model can be used to study this phenomena. Although a wealth of data is available from both human and monkey fMRI studies, fMRI does not have sufficient spatial and temporal resolution to test the responses and selectivity of the neurons. Hence, invasive extracellular single-unit recordings are necessary, which can only be performed in non-human primates. We will train four adult macaque monkeys in passive fixation and depth structure categorization (by means of saccadic eye movements) tasks. We will record extracellular single-cell activities using tungsten microelectrodes. The net neural responses will be calculated by subtracting the mean activity in the 400 ms immediately preceding stimulus onset from the mean activity between 50 and 450 ms after stimulus onset. In every study we will use the minimum number of animals that is necessary for a publication (which is 2), but statistics will be calculated on a large number of neurons (typically 40) in each animal, and multiple trials (typically 10) for each stimulus condition. All data analysis including the statistics and also stimulus generation will be performed using the Matlab (Mathworks) software. We will use a basic set of stimuli consisting of 32 pairs of curved (second order, 4 depth profiles: sine, Gaussian, inclined and S-shape) and 4 pairs of planar surfaces rotated - either along their vertical or horizontal axis - in depth (first order, horizontal and vertical disparity gradients). The stimuli will be presented dichoptically (i.e. different images to the two eyes) by means of a double pair of ferroelectric liquid crystal shutters operating in synchrony with the vertical retrace of a monochrome fast decay (P46) phosphor CRT monitor. Subjects will have to fixate a small target in the center of the display for 400 ms and then the stimulus will be presented for 500 ms, after which the monkey will receive a small amount of juice as reward at the end of stimulus presentation. In both PIP and TEO, we will search for responsive neurons during the presentation of first- and second order stimuli. In responsive neurons, we will test their selectivity for curved and planar surfaces, the responses to monocular presentations and higher-order disparity selectivity by presenting the stimuli at different positions in depth. We will also attempt to map the receptive field of the neurons using the preferred 3D stimulus (presented at 35 positions in the visual field), and investigate size tolerance by presenting the stimuli at a range of sizes (between 4 and 19 deg). Neurons that display clear higher-order disparity selectivity will be tested with a large set of 3D stimuli which represent approximations to the curved surfaces in the search test. We will also test how sensitive these neurons are by presenting curved stimuli at a range of disparity amplitudes (between 0.03 deg and 0.63 deg). In experiment 3, we will first verify the clustering of higher-order disparity selectivity, i.e. neurons within a recording track of 200 to 1000 micron should prefer the same 3D stimulus (e.g. convex). If we observe second-order disparity selectivity and sufficient clustering of the 3D shape preference, we will position the electrode in the center of the recording track, and apply a small amount (35 µAmp) of electrical current in half of the trials while the animal categorizes the depth structure of the stimulus. In this task, the monkey is required to make an eye movement to the left if the surface appears concave and an eye movement to the right if the surface appears convex. A small amount of current will be injected from stimulus onset until the eye movement is detected. We will measure the effect of electrical microstimulation on perceptual choice and on the reaction time of the animal. The results will be compared to those of previous experiments in TEs (Verhoef et al., 2012). Conclusion This project should clarify the nature of responses along two important mid-stage areas in dorsal and ventral visual pathway and their role in depth perception. In addition, it will provide us with valuable and important information about characteristics of responses in the 3D processing network of the brain. Finally, the outcome of this project will be applicable for modeling, brain-machine interfaces and most importantly it will improve our general understanding of one of the most critical and computationally demanding tasks the primate brain carries out every day." "Functional architecture for disparity in macaque inferior temporal cortex and its relationship to the architecture for faces, color, scenes, and visual field" "Bram Verhoef" "Binocular disparity is a powerful depth cue for object perception. The computations for object vision culminate in inferior temporal cortex (IT), but the functional organization for disparity in IT is unknown. Here we addressed this question by measuring fMRI responses in alert monkeys to stimuli that appeared in front of (near), behind (far), or at the fixation plane. We discovered three regions that showed preferential responses for near and far stimuli, relative to zero-disparity stimuli at the fixation plane. These ""near/far"" disparity-biased regions were located within dorsal IT, as predicted by microelectrode studies, and on the posterior inferotemporal gyrus. In a second analysis, we instead compared responses to near stimuli with responses to far stimuli and discovered a separate network of ""near"" disparity-biased regions that extended along the crest of the superior temporal sulcus. We also measured in the same animals fMRI responses to faces, scenes, color, and checkerboard annuli at different visual field eccentricities. Disparity-biased regions defined in either analysis did not show a color bias, suggesting that disparity and color contribute to different computations within IT. Scene-biased regions responded preferentially to near and far stimuli (compared with stimuli without disparity) and had a peripheral visual field bias, whereas face patches had a marked near bias and a central visual field bias. These results support the idea that IT is organized by a coarse eccentricity map, and show that disparity likely contributes to computations associated with both central (face processing) and peripheral (scene processing) visual field biases, but likely does not contribute much to computations within IT that are implicated in processing color." "Robust stereo matching using Census cost, discontinuity-preserving disparity computation and view-consistent refinement" "Minh Duc Nguyen, Jan Hanca, Shaoping Lu" "Stereo matching has been one of the most active research topics in computer vision domain for many years resulting in a large number of techniques proposed in the literature. Nevertheless, improper combinations of available tools cannot fully utilize the advantages of each method and may even lower the performance of the stereo matching system. Moreover, state-of-the-art techniques are usually optimized to perform well on a certain input dataset. In this paper we propose a framework to combine existing tools into a stereo matching pipeline and three different architectures combining existing processing steps to build stereo matching systems which are not only accurate but also efficient and robust under different operating conditions. Thorough experiments on three well-known datasets confirm the effectiveness of our proposed systems on any input data." "Mapping the cause-specific premature mortality reveals large between-districts disparity in Belgium, 2003–2009" "Francoise Therese Renard, J. Tafforeau" "Reducing premature mortality is a crucial public health objective. The goal of this paper, beside updating previous mortality atlases with recent data, is to explore the relative between-districts disparity using a relative-scale map design." "Efficient disparity vector coding for multi-view 3-D displays" "Aykut Avci, Lawrence Bogaert, Roel Beernaert, Jelle De Smet, Youri Meuret, Hugo Thienpont, Herbert De Smet" "Disparity estimation can be used for eliminating redundancies between different views of an object or a scene recorded by an array of cameras which are arranged both horizontally and vertically. However, estimation of the disparity vectors is a highly time consuming process which takes most of the operation time of the multi-view video coding. Therefore, either the amount of data that is to be processed or the complexity of the coding method needs to be decreased in order to encode the multi-view video in a reasonable time. It is proven that the disparities of a point in the scene photographed by cameras which are spaced equidistantly are equal. Since there is a strong geometrical correlation of the disparity vectors, the disparity vector of a view can for most blocks be derived from the disparity vector of another view or views. A new algorithm is presented that reduces the amount of processing time needed for calculating the disparity vectors of each neighboring view except the principal ones. Different schemes are proposed for 3*3 views and they are applied to several image sequences taken from a camera-array. The experimental results show that the proposed schemes yield better results than the reference scheme while preserving the image quality and the amount of encoded data." "Reduced-complexity disparity estimation for efficient multiview imagery encoding" "Aykut Avci" "3D display technology has witnessed a rapid development in the past decades. Currently 3D displays are being widely used in different application areas such as education, broadcasting, entertainment, surgery, video conferencing, etc. However, this technology owes its success to the other complementary technologies such as image acquisition, compression and transmission. In multiview displays, the realism of the reproduced 3D scene is dependent on the number of available views that the displays can show. These view images can be captured from different viewpoints of a scene by using a camera array. A smoother transition between views can be obtained by increasing the number of cameras located in the camera array. However, this comes at the price of an increased amount of image data which needs to be encoded (compressed) to store and transmit the data efficiently. The computational load of the multiview encoder is high and needs to be mitigated. This thesis focuses mainly on encoding multiview images captured by a 2D camera array efficiently in terms of complexity besides quality and bit-rate. The results on this thesis show that the complexity of the encoder can be reduced significantly without compromising the quality and bitrate." "Efficient disparity vector prediction schemes with modified P frame for 2D camera arrays" "Aykut Avci, Jan De Cock, Peter Lambert, Roel Beernaert, Jelle De Smet, Lawrence Bogaert, Youri Meuret, Hugo Thienpont, Herbert De Smet"