Time filter

Source Type

Chouzenoux E.,CNRS Informatics Laboratory of Gaspard Monge Institute | Moussaoui S.,CNRS Research Institute of Communication and Cybernetics of Nantes | Idier J.,CNRS Research Institute of Communication and Cybernetics of Nantes
IEEE Workshop on Statistical Signal Processing Proceedings | Year: 2011

We discuss in this paper the influence of line search on the performances of interior point algorithms applied for constrained signal restoration. Interior point algorithms ensure the fulfillment of the constraints through the minimization of a criterion augmented with a barrier function. However, the presence of the barrier function can slow down the convergence of iterative descent algorithms when general-purpose line search procedures are employed. We recently proposed a line search algorithm, based on a majorization-minimization approach, which allows to handle the singularity introduced by the barrier function. We present here a comparative study of various line search strategies for the resolution of a sparse signal restoration problem with both primal and primal-dual interior point algorithms. © 2011 IEEE.


Chouzenoux E.,CNRS Research Institute of Communication and Cybernetics of Nantes | Chouzenoux E.,CNRS Informatics Laboratory of Gaspard Monge Institute | Moussaoui S.,CNRS Research Institute of Communication and Cybernetics of Nantes | Idier J.,CNRS Research Institute of Communication and Cybernetics of Nantes
Inverse Problems | Year: 2012

This paper focuses on the issue of stepsize determination (linesearch) in iterative descent algorithms applied to the minimization of a criterion containing a barrier function associated with linear constraints. Such an issue arises in inversion methods involving the minimization of a penalized criterion where the barrier function comes either from the data fidelity term or from the regularizing functional. In order to circumvent the inefficiency of general-purpose linesearch strategies in the case of barrier functions, we propose to adopt a majorizationminimization scheme by deriving a new form of a majorant function well suited to approximate a criterion containing barrier terms. We also establish the convergence of classical descent algorithms when this linesearch strategy is employed. Its efficiency is illustrated by means of numerical examples of signal and image restoration. © 2012 IOP Publishing Ltd.


Bojanczyk M.,University of Warsaw | David C.,CNRS Informatics Laboratory of Gaspard Monge Institute | Muscholl A.,University of Bordeaux 1 | Schwentick T.,TU Dortmund | Segoufin L.,Ecole Normale Superieure de Cachan
ACM Transactions on Computational Logic | Year: 2011

In a data word each position carries a label from a finite alphabet and a data value from some infinite domain. This model has been already considered in the realm of semistructured data, timed automata, and extended temporal logics. This article shows that satisfiability for the two-variable fragment FO 2(∼,<,+1) of first-order logic with data equality test ∼ is decidable over finite and infinite data words. Here +1 and < are the usual successor and order predicates, respectively. The satisfiability problem is shown to be at least as hard as reachability in Petri nets. Several extensions of the logic are considered; some remain decidable while some are undecidable. © 2011 ACM.


Mouffranc C.,CNRS Informatics Laboratory of Gaspard Monge Institute | Nozick V.,CNRS Informatics Laboratory of Gaspard Monge Institute
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) | Year: 2013

Colorimetric correction is a necessary task to generate comfortable stereoscopic images. This correction is usually performed with a 3D lookup table that can correct images in real-time and can deal with the non-independence of the colour channels. In this paper, we present a method to compute such 3D lookup table with a non-linear process that minimizes the colorimetric properties of the images. This lookup table is represented by a polynomial basis to reduce the number of required parameters. We also describe some optimizations to speedup the processing time. © 2013 Springer-Verlag.


Salaun Y.,CNRS Informatics Laboratory of Gaspard Monge Institute | Marlet R.,CNRS Informatics Laboratory of Gaspard Monge Institute | Monasse P.,CNRS Informatics Laboratory of Gaspard Monge Institute
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) | Year: 2016

Usual Structure from Motion techniques based on feature points have a hard time on scenes with little texture or presenting a single plane, as in indoor environments. Line segments are more robust features in this case. We propose a novel geometrical criterion for two-view pose estimation using lines, that does not assume a Manhattan world. We also define a parameterless (a contrario) RANSAC-like method to discard calibration outliers and provide more robust pose estimations, possibly using points as well when available. Finally, we provide quantitative experimental data that illustrate failure cases of other methods and that show how our approach outperforms them, both in robustness and precision. © Springer International Publishing AG 2016.


Amano S.,BM Research | David C.,CNRS Informatics Laboratory of Gaspard Monge Institute | Libkin L.,University of Edinburgh | Murlak F.,University of Warsaw
Journal of the ACM | Year: 2014

Relational schema mappings have been extensively studied in connection with data integration and exchange problems, but mappings between XML schemas have not received the same amount of attention. Our goal is to develop a theory of expressive XML schema mappings. Such mappings should be able to use various forms of navigation in a document, and specify conditions on data values. We develop a language for XML schema mappings, and study both data exchange with such mappings and metadata management problems. Specifically, we concentrate on four types of problems: complexity of mappings, query answering, consistency issues, and composition. We first analyze the complexity of mappings, that is, recognizing pairs of documents such that one can be mapped into the other, and provide a classification based on sets of features used in mappings. Next, we chart the tractability frontier for the query answering problem. We show that the problem is tractable for expressive schema mappings and simple queries, but not vice versa. Then, we move to static analysis. We study the complexity of the consistency problem, that is, deciding whether it is possible to map some document of a source schema into a document of the target schema. Finally, we look at composition of XML schema mappings.We analyze its complexity and show that it is harder to achieve closure under composition for XML than for relational mappings. Nevertheless, we find a robust class of XML schema mappings that, in addition to being closed under composition, have good complexity properties with respect to the main data management tasks. Due to its good properties, we suggest this class as the class to use in applications of XML schema mappings. © 2014 ACM.


Mestre X.,Catalonia Technology Center of Telecomunications | Vallet P.,CNRS Informatics Laboratory of Gaspard Monge Institute | Loubaton P.,CNRS Informatics Laboratory of Gaspard Monge Institute | Hachem W.,Telecom ParisTech
IEEE Workshop on Statistical Signal Processing Proceedings | Year: 2011

Traditional estimators of the eigen-subspaces of sample co-variance matrices are known to be consistent only when the sample volume increases for a fixed observation dimension. Due to this fact, their accuracy tends to be rather poor in practical settings where the number of samples and the observation dimension are comparable in magnitude. To overcome this effect, an estimator was recently proposed that provides consistent subspace estimates even when the dimension of the observation scales up with the number of samples. In this paper, the asymptotic distribution of this estimator is characterized by means of a central limit theorem (CLT). © 2011 IEEE.


Vallet P.,CNRS Informatics Laboratory of Gaspard Monge Institute | Vallet P.,Catalonia Technology Center of Telecomunications | Mestre X.,Catalonia Technology Center of Telecomunications | Loubaton P.,CNRS Informatics Laboratory of Gaspard Monge Institute
European Signal Processing Conference | Year: 2012

Subspace methods (e.g. MUSIC) are widely used in the context of DoA estimation using an array of M antennas. These methods perform well as long as the number of available samples N is much larger than M. However, their performance severely degrades when N is of the same order of magnitude than M. In this context, a DoA estimation method (called "G-MUSIC"), based on a new localization function estimate, was recently derived and shown to outperform the traditional methods for reasonable values of M, N. The consistency of both the localization function and DoA estimators was addressed in the asymptotic regime where M, N converge to infinity such that M /N converges to a positive constant. This paper addresses Central Limit Theorems for both the localization function and DoA estimators, in the previous asymptotic regime. Simulations confirm the validity of the results. © 2012 EURASIP.


Boulch A.,ONERA | Marlet R.,CNRS Informatics Laboratory of Gaspard Monge Institute
Computer Graphics Forum | Year: 2016

Normal estimation in point clouds is a crucial first step for numerous algorithms, from surface reconstruction and scene understanding to rendering. A recurrent issue when estimating normals is to make appropriate decisions close to sharp features, not to smooth edges, or when the sampling density is not uniform, to prevent bias. Rather than resorting to manually-designed geometric priors, we propose to learn how to make these decisions, using ground-truth data made from synthetic scenes. For this, we project a discretized Hough space representing normal directions onto a structure amenable to deep learning. The resulting normal estimation method outperforms most of the time the state of the art regarding robustness to outliers, to noise and to point density variation, in the presence of sharp edges, while remaining fast, scaling up to millions of points. © 2016 The Author(s) Computer Graphics Forum © 2016 The Eurographics Association and John Wiley & Sons Ltd. Published by John Wiley & Sons Ltd.


Nozick V.,CNRS Informatics Laboratory of Gaspard Monge Institute
Proceedings of 2011 1st International Symposium on Access Spaces, ISAS 2011 | Year: 2011

This paper presents an extension of image rectification methods for an arbitrary number of views with aligned camera center. This technique can be used for stereoscopic rendering to enhance the perception comfort or for depth from stereo. In this paper, we first expose that epipolar geometry is not suited to solve this problem. Then we propose a non linear method that includes all the images in the rectification process. Our method only requires point correspondences between the views and can handle images with different resolutions. The tests show that the method is robust to noise and and to sparse point correspondences among the views. © 2011 IEEE.

Loading CNRS Informatics Laboratory of Gaspard Monge Institute collaborators
Loading CNRS Informatics Laboratory of Gaspard Monge Institute collaborators