Abstract: Brightness induction is the modulation of the perceived intensity of an area by the luminance of surrounding areas. Recent neurophysiological evidence suggests that brightness information might be explicitly represented in V1, in contrast to the more common assumption that the striate cortex is an area mostly responsive to sensory information. Here we investigate possible neural mechanisms that offer a plausible explanation for such phenomenon. To this end, a neurodynamical model which is based on neurophysiological evidence and focuses on the part of V1 responsible for contextual influences is presented. The proposed computational model successfully accounts for well known psychophysical effects for static contexts and also for brightness induction in dynamic contexts defined by modulating the luminance of surrounding areas. This work suggests that intra-cortical interactions in V1 could, at least partially, explain brightness induction effects and reveals how a common general architecture may account for several different fundamental processes, such as visual saliency and brightness induction, which emerge early in the visual processing pathway.

Abstract: Training a deep architecture using a ranking loss has become standard for the person re-identification task. Increasingly, these deep architectures include additional components that leverage part detections, attribute predictions, pose estimators and other auxiliary information, in order to more effectively localize and align discriminative image regions. In this paper we adopt a different approach and carefully design each component of a simple deep architecture and, critically, the strategy for training it effectively for person re-identification. We extensively evaluate each design choice, leading to a list of good practices for person re-identification. By following these practices, our approach outperforms the state of the art, including more complex methods with auxiliary components, by large margins on four benchmark datasets. We also provide a qualitative analysis of our trained representation which indicates that, while compact, it is able to capture information from localized and discriminative regions, in a manner akin to an implicit attention mechanism.

Abstract: The estimation of the illuminant of a scene from a digital image has been the goal of a large amount of research in computer vision. Color constancy algorithms have dealt with this problem by defining different heuristics to select a unique solution from within the feasible set. The performance of these algorithms has shown that there is still a long way to go to globally solve this problem as a preliminary step in computer vision. In general, performance evaluation has been done by comparing the angular error between the estimated chromaticity and the chromaticity of a canonical illuminant, which is highly dependent on the image dataset. Recently, some workers have used high-level constraints to estimate illuminants; in this case selection is based on increasing the performance on the subsequent steps of the systems. In this paper we propose a new performance measure, the perceptual angular error. It evaluates the performance of a color constancy algorithm according to the perceptual preferences of humans, or naturalness (instead of the actual optimal solution) and is independent of the visual task. We show the results of a new psychophysical experiment comparing solutions from three different color constancy algorithms. Our results show that in more than a half of the judgments the preferred solution is not the one closest to the optimal solution. Our experiments were performed on a new dataset of images acquired with a calibrated camera with an attached neutral grey sphere, which better copes with the illuminant variations of the scene.

Abstract: This paper describes the monomodal and multimodal Neural Machine Translation systems developed by LIUM and CVC for WMT17 Shared Task on Multimodal Translation. We mainly explored two multimodal architectures where either global visual features or convolutional feature maps are integrated in order to benefit from visual context. Our final systems ranked first for both En-De and En-Fr language pairs according to the automatic evaluation metrics METEOR and BLEU.

Abstract: In this paper, computational methods are proposed to compute color edge saliency based on the information content of color edges. The computational methods are evaluated on bottom-up saliency in a psychophysical experiment, and on a more complex task of salient object detection in real-world images. The psychophysical experiment demonstrates the relevance of using information theory as a saliency processing model and that the proposed methods are significantly better in predicting color saliency (with a human-method correspondence up to 74.75% and an observer agreement of 86.8%) than state-of-the-art models. Furthermore, results from salient object detection confirm that an early fusion of color and contrast provide accurate performance to compute visual saliency with a hit rate up to 95.2%.

Abstract: In two previous experiments [Parraga et al, 2009 J. of Im. Sci. and Tech 53(3) 031106; Benavente et al,2009 Perception 38 ECVP Supplement, 36] the boundaries of basic colour categories were measured.
In the first experiment, samples were presented in isolation (ie on a dark background) and boundaries were measured using a yes/no paradigm. In the second, subjects adjusted the chromaticity of a sample presented on a random Mondrian background to find the boundary between pairs of adjacent colours.
Results from these experiments showed significant dierences but it was not possible to conclude whether this discrepancy was due to the absence/presence of a colourful background or to the dierences in the paradigms used. In this work, we settle this question by repeating the first experiment (ie samples presented on a dark background) using the second paradigm. A comparison of results shows that
although boundary locations are very similar, boundaries measured in context are significantly dierent(more diuse) than those measured in isolation (confirmed by a Student’s t-test analysis on the subject’s answers statistical distributions). In addition, we completed the mapping of colour name space by measuring the boundaries between chromatic colours and the achromatic centre. With these results we
completed our parametric fuzzy-sets model of colour naming space.

Abstract: The color appearance of a surface depends on the color of its surroundings (inducers). When the perceived color shifts towards that of the surroundings, the effect is called “color assimilation” and when it shifts away from the surroundings it is called “color contrast.” There is also evidence that the phenomenon depends on the spatial configuration of the inducer, e.g., uniform surrounds tend to induce color contrast and striped surrounds tend to induce color assimilation. However, previous work found that striped surrounds under certain conditions do not induce color assimilation but induce color contrast (or do not induce anything at all), suggesting that luminance differences and high spatial frequencies could be key factors in color assimilation. Here we present a new psychophysical study of color assimilation where we assessed the contribution of luminance differences (between the target and its surround) present in striped stimuli. Our results show that luminance differences are key factors in color assimilation for stimuli varying along the s axis of MacLeod-Boynton color space, but not for stimuli varying along the l axis. This asymmetry suggests that koniocellular neural mechanisms responsible for color assimilation only contribute when there is a luminance difference, supporting the idea that mutual-inhibition has a major role in color induction.

Abstract: Brightness induction is the modulation of the perceived intensity of an area by the luminance of surrounding areas and reveals fundamental properties of neural organization in the visual system. Several phenomenological models have been proposed that successfully account for psychophysical data (Pessoa et al. 1995, Blakeslee and McCourt 2004, Barkan et al. 2008, Otazu et al. 2008).
Neurophysiological evidence suggests that brightness information is explicitly represented in V1 and neuronal response modulations have been observed followingluminance changes outside their receptive fields (Rossi and Paradiso, 1999).
In this work we investigate possible neural mechanisms that offer a plausible explanation for such effects. To this end, we consider the model by Z.Li (1999) which is based on biological data and focuses on the part of V1 responsible for contextual influences, namely, layer 2–3 pyramidal cells, interneurons, and horizontal intracortical connections. This model has proven to account for phenomena such as contour detection and preattentive segmentation, which share with brightness induction the relevant effect of contextual influences. In our model, the input to the network is derived from a complete multiscale and multiorientation wavelet decomposition which makes it possible to recover an image reflecting the perceived intensity. The proposed model successfully accounts for well known pyschophysical effects (among them: the White's and modified White's effects, the Todorović, Chevreul, achromatic ring patterns, and grating induction effects). Our work suggests that intra-cortical interactions in the primary visual cortex could partially explain perceptual brightness induction effects and reveals how a common general architecture may account for several different fundamental processes emerging early in the visual pathway.

Abstract: Recognizing materials and textures in realistic imaging conditions is a challenging computer vision problem. For many years, local features based orderless representations were a dominant approach for texture recognition. Recently deep local features, extracted from the intermediate layers of a Convolutional Neural Network (CNN), are used as filter banks. These dense local descriptors from a deep model, when encoded with Fisher Vectors, have shown to provide excellent results for texture recognition. The CNN models, employed in such approaches, take RGB patches as input and train on a large amount of labeled images. We show that CNN models, which we call TEX-Nets, trained using mapped coded images with explicit texture information provide complementary information to the standard deep models trained on RGB patches. We further investigate two deep architectures, namely early and late fusion, to combine the texture and color information. Experiments on benchmark texture datasets clearly demonstrate that TEX-Nets provide complementary information to standard RGB deep network. Our approach provides a large gain of 4.8%, 3.5%, 2.6% and 4.1% respectively in accuracy on the DTD, KTH-TIPS-2a, KTH-TIPS-2b and Texture-10 datasets, compared to the standard RGB network of the same architecture. Further, our final combination leads to consistent improvements over the state-of-the-art on all four datasets.