Search

This paper presents adaptive subspace-based inverse projections via division into multiple sub-problems (ASIP-DIMSs) for missing image data restoration. In the proposed method, a target problem for estimating missing image data is divided into multiple sub-problems, and each sub-problem is iteratively solved with the constraints of other known image data. By projection into a subspace model of image patches, the solution of each sub-problem is calculated, where we call this procedure “subspace-based inverse projection” for simplicity. The proposed method can use higher dimensional subspaces for finding unique solutions in each sub-problem, and successful restoration becomes feasible, since a high level of image representation performance can be preserved. This is the biggest contribution of this paper. Furthermore, the proposed method generates several subspaces from known training examples and enables derivation of a new criterion in the above framework to adaptively select the optimal subspace for each target patch. In this way, the proposed method realizes missing image data restoration using ASIP-DIMS. Since our method can estimate any kind of missing image data, its potential in two image restoration tasks, image inpainting and super-resolution, based on several methods for multivariate analysis is also shown in this paper.

This paper studies a new Bayesian unmixing algorithm for hyperspectral images. Each pixel of the image is modeled as a linear combination of so-called endmembers. These endmembers are supposed to be random in order to model uncertainties regarding their knowledge. More precisely, we model endmembers as Gaussian vectors whose means have been determined using an endmember extraction algorithm such as the famous N-finder (N-FINDR) or Vertex Component Analysis (VCA) algorithms. This paper proposes to estimate the mixture coefficients (referred to as abundances) using a Bayesian algorithm. Suitable priors are assigned to the abundances in order to satisfy positivity and additivity constraints whereas conjugate priors are chosen for the remaining parameters. A hybrid Gibbs sampler is then constructed to generate abundance and variance samples distributed according to the joint posterior of the abundances and noise variances. The performance of the proposed methodology is evaluated by comparison with other unmixing algorithms on synthetic and real images.

International audience; This paper studies a family of distributions constructed from multivariate gamma distributions to model the statistical properties of multisensor synthetic aperture radar (SAR) images. These distributions referred to as multisensor multivariate gamma distributions (MuMGDs) are potentially interesting for detecting changes in SAR images acquired by different sensors having different numbers of looks. The first part of the paper compares different estimators for the parameters of MuMGDs. These estimators are based on the maximum likelihood principle, the method of inference function for margins and the method of moments. The second part of the paper studies change detection algorithms based on the estimated correlation coefficient of MuMGDs. Simulation results conducted on synthetic and real data illustrate the performance of these change detectors.

Tree-like structures such as retinal images are widely studied in computer-aided diagnosis systems for large-scale screening programs. Despite several segmentation and tracking methods proposed in the literature, there still exist several limitations specifically when two or more curvilinear structures cross or bifurcate, or in the presence of interrupted lines or highly curved blood vessels. In this paper, we propose a novel approach based on multi-orientation scores augmented with a contextual affinity matrix, which both are inspired by the geometry of the primary visual cortex (V1) and their contextual connections. The connectivity is described with a five-dimensional kernel obtained as the fundamental solution of the Fokker-Planck equation modelling the cortical connectivity in the lifted space of positions, orientations, curvatures and intensity. It is further used in a self-tuning spectral clustering step to identify the main perceptual units in the stimuli. The proposed method has been validated on several easy and challenging structures in a set of artificial images and actual retinal patches. Supported by quantitative and qualitative results, the method is capable of overcoming the limitations of current state-of-the-art techniques.

In this paper, we present a multichannel cross-modal fusion algorithm to combine two complementary modalities in electron tomography: X-ray spectroscopy and scanning transmission electron microscopy (STEM). The former reveals compositions with high elemental specificity but low signal-to-noise ratio (SNR), while the latter characterizes the structure with high SNR but little chemical information. We use a multivariate regression to build a cross-modal fusion framework for these two modalities to simultaneously achieve high elemental specificity and high SNR for a target element chosen from the sample under study. Specifically, we first compute three-dimensional tomograms from tilt-series datasets of X-ray and STEM using different reconstruction algorithms. Then, we generate many feature images from each tomogram. Finally, we adopt partial least squares regression to assess the connection between these feature images and the reconstruction of the target element. Based on the simulated and experimental datasets of semiconductor devices, we demonstrate that our algorithm can not only produce continuous edges, homogeneous foreground, and clean background in its element-specific reconstructions but also can more accurately preserve fine structures than state-of-the-art tomography techniques. Moreover, we show that it can deliver results with high fidelity even for X-ray datasets with limited tilts or low counts. This property is highly desired in the semiconductor industry where acquisition time and sample damage are essential.

Magnetic resonance Diffusion Tensor Imaging (DTI) measures diffusion of water molecules and is used to characterize orientation of white matter fibers and connectivity of neurological structures. Segmentation and visualization of DT images is challenging, because of low data quality and complexity of anatomical structures. In this paper, we propose an interactive segmentation approach, based on a hierarchical representation of the input DT image through a tree structure. The tree is obtained by successively merging watershed regions, based on the morphological waterfall approach, hence the name watershed tree. Region merging is done according to a combined similarity and homogeneity criterion. We introduce filters that work on the proposed tree representation, and that enable region-based attribute filtering of DTI data. Linked views between the visualizations of the simplified DT image and the tree enable a user to visually explore both data and tree at interactive rates. The coupling of filtering, semi-automatic segmentation by labeling nodes in the tree, and various interaction mechanisms support the segmentation task. Our method is robust against noise, which we demonstrate on synthetic and real DTI data. Keywords: DTI; multiple views; segmentation; waterfall algorithm; watershed

The direction of sunlight is an important informative cue in a number of applications in image processing, such as augmented reality and object recognition. In general, existing methods to estimate the direction of the sunlight rely on different image features (e.g., sky, texture, shadows, and shading). These features can be considered as weak informative cues as no single feature can reliably estimate the sunlight direction. Moreover, existing methods may require that the camera parameters are known limiting their applicability. In this paper, we present a new method to estimate the sunlight direction from a single (outdoor) image by inferring casts shadows through object modeling and recognition. First, objects (e.g., cars or persons) are first (automatically) recognized in images by exemplar-SVMs. Instead of training the Support Vector Machine (SVMs) using natural images (limited variation in viewpoints), we propose to train on 2D object samples generated from 3D object models. Then, the recognized objects are used as sundial cues (probes) to estimate the sunlight direction by inferring the corresponding shadows generated by 3D object models considering different illumination directions. We demonstrate the effectiveness of our approach on synthetic and real images. Experiments show that our method estimates the azimuth angle accurately within a quadrant (smaller than 45°) and compute the zenith angle with mean angular error of 23°.

Extracting the pixel-level 3D layout from a single image is important for different applications, such as object localization, image, and video categorization. Traditionally, the 3D layout is derived by solving a pixel-level classification problem. However, the image-level 3D structure can be very beneficial for extracting pixel-level 3D layout since it implies the way how pixels in the image are organized. In this paper, we propose an approach that first predicts the global image structure, and then we use the global structure for fine-grained pixel-level 3D layout extraction. In particular, image features are extracted based on multiple layout templates. We then learn a discriminative model for classifying the global layout at the image-level. Using latent variables, we implicitly model the sublevel semantics of the image, which enrich the expressiveness of our model. After the image-level structure is obtained, it is used as the prior knowledge to infer pixel-wise 3D layout. Experiments show that the results of our model outperform the state-of-the-art methods by 11.7% for 3D structure classification. Moreover, we show that employing the 3D structure prior information yields accurate 3D scene layout segmentation.

The additive white Gaussian noise is widely assumed in many image processing algorithms. However, in the real world, the noise from actual cameras is better modeled as signal-dependent noise (SDN). In this paper, we focus on the SDN model and propose an algorithm to automatically estimate its parameters from a single noisy image. The proposed algorithm identifies the noise level function of signal-dependent noise assuming the generalized signal-dependent noise model and is also applicable to the Poisson-Gaussian noise model. The accuracy is achieved by improved estimation of local mean and local noise variance from the selected low-rank patches. We evaluate the proposed algorithm with both synthetic and real noisy images. Experiments demonstrate that the proposed estimation algorithm outperforms the state-of-the-art methods.

The classification of retinal vessels into artery/vein (A/V) is an important phase for automating the detection of vascular changes, and for the calculation of characteristic signs associated with several systemic diseases such as diabetes, hypertension, and other cardiovascular conditions. This paper presents an automatic approach for A/V classification based on the analysis of a graph extracted from the retinal vasculature. The proposed method classifies the entire vascular tree deciding on the type of each intersection point (graph nodes) and assigning one of two labels to each vessel segment (graph links). Final classification of a vessel segment as A/V is performed through the combination of the graph-based labeling results with a set of intensity features. The results of this proposed method are compared with manual labeling for three public databases. Accuracy values of 88.3%, 87.4%, and 89.8% are obtained for the images of the INSPIRE-AVR, DRIVE, and VICAVR databases, respectively. These results demonstrate that our method outperforms recent approaches for A/V classification.

Many computer vision applications, including image classification, matching, and retrieval use global image representations, such as the Fisher vector, to encode a set of local image patches. To describe these patches, many local descriptors have been designed to be robust against lighting changes and noise. However, local image descriptors are unstable when the underlying image signal is low. Such low-signal patches are sensitive to small image perturbations, which might come e.g., from camera noise or lighting effects. In this paper, we first quantify the relation between the signal strength of a patch and the instability of that patch, and second, we extend the standard Fisher vector framework to explicitly take the descriptor instabilities into account. In comparison to common approaches to dealing with descriptor instabilities, our results show that modeling local descriptor instability is beneficial for object matching, image retrieval, and classification.

The aim of color constancy is to remove the effect of the color of the light source. As color constancy is inherently an ill-posed problem, most of the existing color constancy algorithms are based on specific imaging assumptions (e.g., gray-world and white patch assumption). In this paper, 3D geometry models are used to determine which color constancy method to use for the different geometrical regions (depth/layer) found in images. The aim is to classify images into stages (rough 3D geometry models). According to stage models, images are divided into stage regions using hard and soft segmentation. After that, the best color constancy methods are selected for each geometry depth. To this end, we propose a method to combine color constancy algorithms by investigating the relation between depth, local image statistics, and color constancy. Image statistics are then exploited per depth to select the proper color constancy method. Our approach opens the possibility to estimate multiple illuminations by distinguishing nearby light source from distant illuminations. Experiments on state-of-the-art data sets show that the proposed algorithm outperforms state-of-the-art single color constancy algorithms with an improvement of almost 50% of median angular error. When using a perfect classifier (i.e, all of the test images are correctly classified into stages); the performance of the proposed method achieves an improvement of 52% of the median angular error compared with the best-performing single color constancy algorithm.

A dynamic texture (DT) is an extension of the texture to the temporal domain. How to segment a DT is a challenging problem. In this paper, we address the problem of segmenting a DT into disjoint regions. A DT might be different from its spatial mode (i.e., appearance) and/or temporal mode (i.e., motion field). To this end, we develop a framework based on the appearance and motion modes. For the appearance mode, we use a new local spatial texture descriptor to describe the spatial mode of the DT; for the motion mode, we use the optical flow and the local temporal texture descriptor to represent the temporal variations of the DT. In addition, for the optical flow, we use the histogram of oriented optical flow (HOOF) to organize them. To compute the distance between two HOOFs, we develop a simple effective and efficient distance measure based on Weber's law. Furthermore, we also address the problem of threshold selection by proposing a method for determining thresholds for the segmentation method by an offline supervised statistical learning. The experimental results show that our method provides very good segmentation results compared to the state-of-the-art methods in segmenting regions that differ in their dynamics.

Complexities of dynamic volumetric imaging challenge the available computer vision techniques on a number of different fronts. This paper examines the relationship between the estimation accuracy and required amount of smoothness for a general solution from a robust statistics perspective. We show that a (surprisingly) small amount of local smoothing is required to satisfy both the necessary and sufficient conditions for accurate optic flow estimation. This notion is called 'just enough' smoothing, and its proper implementation has a profound effect on the preservation of local information in processing 3D dynamic scans. To demonstrate the effect of 'just enough' smoothing, a robust 3D optic flow method with quantized local smoothing is presented, and the effect of local smoothing on the accuracy of motion estimation in dynamic lung CT images is examined using both synthetic and real image sequences with ground truth.

International audience; This paper proposes a joint segmentation and deconvolution Bayesian method for medical ultrasound (US) images. Contrary to piecewise homogeneous images, US images exhibit heavy characteristic speckle patterns correlated with the tissue structures. The generalized Gaussian distribution (GGD) has been shown to be one of the most relevant distributions for characterizing the speckle in US images. Thus, we propose a GGD-Potts model defined by a label map coupling US image segmentation and deconvolution. The Bayesian estimators of the unknown model parameters, including the US image, the label map, and all the hyperparameters are difficult to be expressed in a closed form. Thus, we investigate a Gibbs sampler to generate samples distributed according to the posterior of interest. These generated samples are finally used to compute the Bayesian estimators of the unknown parameters. The performance of the proposed Bayesian model is compared with the existing approaches via several experiments conducted on realistic synthetic data and in vivo US images.

Head pose and eye location for gaze estimation have been separately studied in numerous works in the literature. Previous research shows that satisfactory accuracy in head pose and eye location estimation can be achieved in constrained settings. However, in the presence of nonfrontal faces, eye locators are not adequate to accurately locate the center of the eyes. On the other hand, head pose estimation techniques are able to deal with these conditions; hence, they may be suited to enhance the accuracy of eye localization. Therefore, in this paper, a hybrid scheme is proposed to combine head pose and eye location information to obtain enhanced gaze estimation. To this end, the transformation matrix obtained from the head pose is used to normalize the eye regions, and in turn, the transformation matrix generated by the found eye location is used to correct the pose estimation procedure. The scheme is designed to enhance the accuracy of eye location estimations, particularly in low-resolution videos, to extend the operative range of the eye locators, and to improve the accuracy of the head pose tracker. These enhanced estimations are then combined to obtain a novel visual gaze estimation system, which uses both eye location and head information to refine the gaze estimates. From the experimental results, it can be derived that the proposed unified scheme improves the accuracy of eye estimations by 16% to 23%. Furthermore, it considerably extends its operating range by more than 15° by overcoming the problems introduced by extreme head poses. Moreover, the accuracy of the head pose tracker is improved by 12% to 24%. Finally, the experimentation on the proposed combined gaze estimation system shows that it is accurate (with a mean error between 2° and 5°) and that it can be used in cases where classic approaches would fail without imposing restraints on the position of the head.

Color constancy algorithms are generally based on the simplifying assumption that the spectral distribution of a light source is uniform across scenes. However, in reality, this assumption is often violated due to the presence of multiple light sources. In this paper, we will address more realistic scenarios where the uniform light-source assumption is too restrictive. First, a methodology is proposed to extend existing algorithms by applying color constancy locally to image patches, rather than globally to the entire image. After local (patch-based) illuminant estimation, these estimates are combined into more robust estimations, and a local correction is applied based on a modified diagonal model. Quantitative and qualitative experiments on spectral and real images show that the proposed methodology reduces the influence of two light sources simultaneously present in one scene. If the chromatic difference between these two illuminants is more than 1° , the proposed framework outperforms algorithms based on the uniform light-source assumption (with error-reduction up to approximately 30%). Otherwise, when the chromatic difference is less than 1° and the scene can be considered to contain one (approximately) uniform light source, the performance of the proposed method framework is similar to global color constancy methods.

International audience; Mixing phenomena in hyperspectral images depend on a variety of factors, such as the resolution of observation devices, the properties of materials, and how these materials interact with incident light in the scene. Different parametric and nonparametric models have been considered to address hyperspectral unmixing problems. The simplest one is the linear mixing model. Nevertheless, it has been recognized that the mixing phenomena can also be nonlinear. The corresponding nonlinear analysis techniques are necessarily more challenging and complex than those employed for linear unmixing. Within this context, it makes sense to detect the nonlinearly mixed pixels in an image prior to its analysis, and then employ the simplest possible unmixing technique to analyze each pixel. In this paper, we propose a technique for detecting nonlinearly mixed pixels. The detection approach is based on the comparison of the reconstruction errors using both a Gaussian process regression model and a linear regression model. The two errors are combined into a detection statistics for which a probability density function can be reasonably approximated. We also propose an iterative endmember extraction algorithm to be employed in combination with the detection algorithm. The proposed detect-then-unmix strategy, which consists of extracting endmembers, detecting nonlinearly mixed pixels and unmixing, is tested with synthetic and real images.

Computational color constancy is a fundamental prerequisite for many computer vision applications. This paper presents a survey of many recent developments and state-of-the-art methods. Several criteria are proposed that are used to assess the approaches. A taxonomy of existing algorithms is proposed and methods are separated in three groups: static methods, gamut-based methods, and learning-based methods. Further, the experimental setup is discussed including an overview of publicly available datasets. Finally, various freely available methods, of which some are considered to be state of the art, are evaluated on two datasets.

Image reconstruction from its projections is a necessity in many applications such as medical (CT), security, inspection, and others. This paper extends the 2-D Fan-beam method in [2] to 3-D. The algorithm, called Pyramid Beam (PB), is based upon the parallel reconstruction algorithm in [1]. It allows fast capturing of the scanned data, and in 3-D, the reconstructions are based upon the discrete X-ray transform [1]. The PB geometries are reordered to fit parallel projection geometry. The underlying idea is to use the algorithm in [1] by porting the proposed PB geometries to fit the algorithm in [1]. The complexity of the algorithm is comparable with the 3-D FFT. The results show excellent reconstruction qualities while being simple for practical use.

An efficient approach toward a no-reference ringing metric intrinsically exists of two steps: first detecting regions in an image where ringing might occur, and second quantifying the ringing annoyance in these regions. This paper presents a novel approach toward the first step: the automatic detection of regions visually impaired by ringing artifacts in compressed images. It is a no-reference approach, taking into account the specific physical structure of ringing artifacts combined with properties of the human visual system (HVS). To maintain low complexity for real-time applications, the proposed approach adopts a perceptually relevant edge detector to capture regions in the image susceptible to ringing, and a simple yet efficient model of visual masking to determine ringing visibility. The approach is validated with the results of a psychovisual experiment, and its performance is compared to existing alternatives in literature for ringing region detection. Experimental results show that our method is promising in terms of both reliability and computational efficiency. © 2010 IEEE.

The Max-Tree designed for morphological attribute filtering in image processing, is a data structure in which the nodes represent connected components for all threshold levels in a data set. Attribute filters compute some attribute describing the shape or size of each connected component and then decide which components to keep or to discard. In this paper, we augment the basic Max-Tree data structure such that interactive volumetric filtering and visualization becomes possible. We introduce extensions that allow 1) direct, splatting-based, volume rendering; 2) representation of the Max-Tree on graphics hardware; and 3) fast active cell selection for isosurface generation. In all three cases, we can use the Max-Tree representation for visualization directly, without needing to reconstruct the volumetric data explicitly. We show that both filtering and visualization can be performed at interactive frame rates, ranging between 2.4 and 32 frames per seconds. In contrast, a standard texture-based volume visualization method manages only between 0.5 and 1.8 frames per second. For isovalue browsing, the experimental results show that the performance is comparable to the performance of an interval tree, where our method has the advantage that both filter threshold browsing and isolevel browsing are fast. It is shown that the methods using graphics hardware can be extended to other connected filters.

Two important visual properties of paintings and painting-like images are the absence of texture details and the increased sharpness of edges as compared to photographic images. Painting-like artistic effects can be achieved from photographic images by filters that smooth out texture details, while preserving or enhancing edges and corners. However, not all edge preserving smoothers are suitable for this purpose. We present a simple nonlinear local operator that generalizes both the well known Kuwahara filter and the more general class of filters known in the literature as ldquocriterion and value filter structure.rdquo This class of operators suffers from intrinsic theoretical limitations which give rise to a dramatic instability in presence of noise, especially on shadowed areas. Such limitations are discussed in the paper and overcome by the proposed operator. A large variety of experimental results shows that the output of the proposed operator is visually similar to a painting. Comparisons with existing techniques on a large set of natural images highlight conditions on which traditional edge preserving smoothers fail, whereas our approach produces good results. In particular, unlike many other well established approaches, the proposed operator is robust to degradations of the input image such as blurring and noise contamination.

Homomorphic property of public-key cryptosystems is applied for several cryptographic protocols, such as electronic cash, voting system, bidding protocols, etc. Several fingerprinting protocols also exploit the property to achieve an asymmetric system. However, their enciphering rate is extremely low and the implementation of watermarking technique is difficult. In this paper, we propose a new fingerprinting protocol applying additive homomorphic property of Okamoto-Uchiyama encryption scheme. Exploiting the property ingenuously, the enciphering rate of our fingerprinting scheme can be close to the corresponding cryptosystem. We study the problem of implementation of watermarking technique and propose a successful method to embed an encrypted information without knowing the plain value. The security can also be protected for both a buyer and a merchant in our scheme.

X-ray computed tomography (CT) is a powerful tool for noninvasive imaging of time-varying objects. In the past, methods have been proposed to reconstruct images from continuously changing objects. For discretely or structurally changing objects, however, such methods fail to reconstruct high quality images, mainly because assumptions about continuity are no longer valid. In this paper, we propose a method to reconstruct structurally changing objects. Starting from the observation that there exist regions within the scanned object that remain unchanged over time, we introduce an iterative optimization routine that can automatically determine these regions and incorporate this knowledge in an algebraic reconstruction method. The proposed algorithm was validated on simulation data and experimental mu CT data, illustrating its capability to reconstruct structurally changing objects more accurately in comparison to current techniques.

Filtered backprojection, one of the most widely used reconstruction methods in tomography, requires a large number of low-noise projections to yield accurate reconstructions. In many applications of tomography, complete projection data of high quality cannot be obtained, because of practical considerations. Algebraic methods tend to handle such problems better, but are computationally more expensive. In this paper, we introduce a new method that improves the filtered backprojection method by using a custom data-dependent filter that minimizes the projection error of the resulting reconstruction. We show that the computational cost of the new method is significantly lower than that of algebraic methods. Experiments on both simulation and experimental data show that the method is able to produce more accurate reconstructions than filtered backprojection based on popular static filters when presented with data with a limited number of projections or statistical noise present. Furthermore, the results show that the method produces reconstructions with similar accuracy to algebraic methods, but is faster at producing them. Finally, we show that the method can be extended to exploit certain forms of prior knowledge, improving reconstruction accuracy in specific cases.

Denoising algorithms based on wavelet thresholding replace small wavelet coefficients by zero and keep or shrink the coefficients with absolute value above the threshold. The optimal threshold minimizes the error of the result as compared to the unknown, exact data. To estimate this optimal threshold, we use generalized cross validation. This procedure does not require an estimation for the noise energy. Originally, this method assumes uncorrelated noise. In this paper, we describe how we can extend it to images with correlated noise. ispartof: IEEE transactions on image processing vol:8 issue:7 pages:947-953 ispartof: location:United States status: published

This paper presents a nonlinear mixing model for hyperspectral image unmixing. The proposed model assumes that the pixel reflectances are nonlinear functions of pure spectral components contaminated by an additive white Gaussian noise. These nonlinear functions are approximated using polynomial functions leading to a polynomial postnonlinear mixing model. A Bayesian algorithm and optimization methods are proposed to estimate the parameters involved in the model. The performance of the unmixing strategies is evaluated by simulations conducted on synthetic and real data.

Hyperspectral images present some specific characteristics that should be used by an efficient compression system. In compression, wavelets have shown a good adaptability to a wide range of data, while being of reasonable complexity. Some wavelet-based compression algorithms have been successfully used for some hyperspectral space missions. This paper focuses on the optimization of a full wavelet compression system for hyperspectral images. Each step of the compression algorithm is studied and optimized. First, an algorithm to find the optimal 3-D wavelet decomposition in a rate-distortion sense is defined. Then, it is shown that a specific fixed decomposition has almost the same performance, while being more useful in terms of complexity issues. It is shown that this decomposition significantly improves the classical isotropic decomposition. One of the most useful properties of this fixed decomposition is that it allows the use of zero tree algorithms. Various tree structures, creating a relationship between coefficients, are compared. Two efficient compression methods based on zerotree coding (EZW and SPIHT) are adapted on this near-optimal decomposition with the best tree structure found. Performances are compared with the adaptation of JPEG 2000 for hyperspectral images on six different areas presenting different statistical properties.

In this paper, a Bayesian wavelet-based denoising procedure for multicomponent images is proposed. A denoising procedure is constructed that (1) fully accounts for the multicomponent image covariances, (2) makes use of Gaussian scale mixtures as prior models that approximate the marginal distributions of the wavelet coefficients well, and (3) makes use of a noise-free image as extra prior information. It is shown that such prior information is available with specific multicomponent image data of, e.g., remote sensing and biomedical imaging. Experiments are conducted in these two domains, in both simulated and real noisy conditions.

In this paper, a denoising technique for multivalued images exploiting interband correlations is proposed. A redundant wavelet transform is applied and denoising is applied by thresholding wavelet coefficients. Specific functions of the wavelet coefficients are defined that exploit interscale and/or interband correlation of the signal. Three functions are studied: the square of the wavelet coefficients, products of coefficients at adjacent scales, and products of coefficients from different bands. For these functions, the signal and noise probability density functions (pdf) become more separated. The high signal correlation between bands is exploited by summing these products over all bands, in this way separating noise and signal pdfs even more. The noise pdf of the proposed quantities is derived analytically and from this, a wavelet threshold is derived. The technique is demonstrated to outperform single band wavelet thresholding on multispectral remote sensing images and on multimodal MRI images.

Removing reflection artefacts from a single image is a problem of both theoretical and practical interest, which still presents challenges because of the massively ill-posed nature of the problem. In this paper, we propose a technique based on a novel optimization problem. First, we introduce a simple user interaction scheme, which helps minimize information loss in the reflection-free regions. Second, we introduce an H 2 fidelity term, which preserves fine detail while enforcing the global color similarity. We show that this combination allows us to mitigate the shortcomings in structure and color preservation, which presents some of the most prominent drawbacks in the existing methods for reflection removal. We demonstrate, through numerical and visual experiments, that our method is able to outperform the state-of-the-art model-based methods and compete with recent deep-learning approaches.

In this paper, we present a new algorithm for fast, online 3D reconstruction of dynamic scenes using times of arrival of photons recorded by single-photon detector arrays. One of the main challenges in 3D imaging using single-photon lidar in practical applications is the presence of strong ambient illumination which corrupts the data and can jeopardize the detection of peaks/surface in the signals. This background noise not only complicates the observation model classically used for 3D reconstruction but also the estimation procedure which requires iterative methods. In this work, we consider a new similarity measure for robust depth estimation, which allows us to use a simple observation model and a non-iterative estimation procedure while being robust to mis-specification of the background illumination model. This choice leads to a computationally attractive depth estimation procedure without significant degradation of the reconstruction performance. This new depth estimation procedure is coupled with a spatio-temporal model to capture the natural correlation between neighboring pixels and successive frames for dynamic scene analysis. The resulting online inference process is scalable and well suited for parallel implementation. The benefits of the proposed method are demonstrated through a series of experiments conducted with simulated and real single-photon lidar videos, allowing the analysis of dynamic scenes at 325 m observed under extreme ambient illumination conditions., 12 pages

Learning discriminative, view-invariant and multi-scale representations of object appearance with different semantic levels is of paramount importance for person Re-Identification (ReID). Recently, the community has focused on learning deep Re-ID models to capture a single holistic representation. To improve the achieved results, additional visual attributes and object part-driven models have been considered, inevitably introducing additional human annotation labor or computational efforts. In this paper, we argue that pyramid-inspired methods capturing multi-scale information may overcome such requirements. Precisely, multi-scale pooled regions representing visual information of an object are integrated within a novel deep architecture factorizing them into discriminative features at multiple semantic levels. These are exploited through an attention mechanism later considered in an identification-similarity multi-task loss, trained by means of a curriculum learning strategy. Extensive results on three person ReID benchmarks demonstrate that better performance than existing methods are achieved. Code is available at https://github.com/iN1k1 .

The goal of psychometric scaling is the quantification of perceptual experiences, understanding the relationship\ud between an external stimulus, the internal representation and the\ud response. In this paper, we propose a probabilistic framework to\ud fuse the outcome of different psychophysical experimental protocols, namely rating and pairwise comparisons experiments. Such\ud a method can be used for merging existing datasets of subjective\ud nature and for experiments in which both measurements are\ud collected. We analyze and compare the outcomes of both types\ud of experimental protocols in terms of time and accuracy in a\ud set of simulations and experiments with benchmark and realworld image quality assessment datasets, showing the necessity of\ud scaling and the advantages of each protocol and mixing. Although\ud most of our examples focus on image quality assessment, our\ud findings generalize to any other subjective quality-of-experience\ud task.

Face images convey rich information which can be perceived as a superposition of low-complexity components associated with attributes, such as facial identity, expressions, and activation of facial action units (AUs). For instance, low-rank components characterizing neutral facial images are associated with identity, while sparse components capturing non-rigid deformations occurring in certain face regions reveal expressions and AU activations. In this paper, the discriminant incoherent component analysis (DICA) is proposed in order to extract low-complexity components, corresponding to facial attributes, which are mutually incoherent among different classes (e.g., identity, expression, and AU activation) from training data, even in the presence of gross sparse errors. To this end, a suitable optimization problem, involving the minimization of nuclear-and $\ell _{1}$ -norm, is solved. Having found an ensemble of class-specific incoherent components by the DICA, an unseen (test) image is expressed as a group-sparse linear combination of these components, where the non-zero coefficients reveal the class(es) of the respective facial attribute(s) that it belongs to. The performance of the DICA is experimentally assessed on both synthetic and real-world data. Emphasis is placed on face analysis tasks, namely, joint face and expression recognition, face recognition under varying percentages of training data corruption, subject-independent expression recognition, and AU detection by conducting experiments on four data sets. The proposed method outperforms all the methods that are compared with all the tasks and experimental settings.