Les récentes évolutions des technologies de l'information et de la communication font qu'il est aujourd'hui facile de consulter des catalogues de morceaux de musique conséquents. De nouvelles représentations et de nouveaux algorithmes doivent de ce fait être développés afin de disposer d'une vision représentative de ces catalogues et de naviguer avec agilité dans leurs contenus. Ceci nécessite une caractérisation efficace des morceaux de musique par l'intermédiaire de descriptions macroscopiques pertinentes. Dans cette thèse, nous nous focalisons sur l'estimation de la structure des morceaux de musique : il s'agit de produire pour chaque morceau une description de son organisation par une séquence de quelques dizaines de segments structurels, définis par leurs frontières (un instant de début et un instant de fin) et par une étiquette représentant leur contenu sonore.La notion de structure musicale peut correspondre à de multiples acceptions selon les ...

L¢objectif de cette these consiste en l¢etude et la construction de nouvelles transformees scalables mises en jeu dans le schema de codage video t+2D, afin d¢ameliorer le gain en compression. L¢utilisation du formalisme lifting lors de la construction de ces transformees spatio-temporelles permet l¢introduction d¢operateurs non-lineaires, particulierement utiles pour representer efficacement les singularites et discontinuites presentes dans une sequence video. Nous nous interessons dans un premier temps a l¢optimisation et a la construction de nouvelles transformees temporelles compensees en mouvement, afin d¢augmenter l¢efficacite de codage objective et subjective. Nous envisageons ensuite l¢elaboration et la mise en place de bancs de filtres M-bandes pour decomposer spatialement les sous-bande temporelles. Nous traitons alors de l¢extension des proprietes de scalabilite du banc de synthese M-bandes a des facteurs rationnels quelconques. Enfin, nous decrivons la construction de decompositions spatiales en ondelettes adaptatives, non-lineaires et ...

In French. See thesis pdf.

Matrices, as natural representation of linear mappings in finite dimension, play a crucial role in signal processing and machine learning. Multiplying a vector by a full rank matrix a priori costs of the order of the number of non-zero entries in the matrix, in terms of arithmetic operations. However, matrices exist that can be applied much faster, this property being crucial to the success of certain linear transformations, such as the Fourier transform or the wavelet transform. What is the property that allows these matrices to be applied rapidly ? Is it easy to verify ? Can weapproximate matrices with ones having this property ? Can we estimate matrices having this property ? This thesis investigates these questions, exploring applications such as learning dictionaries with efficient implementations, accelerating the resolution of inverse problems or Fast Fourier Transform on graphs.

An intelligent antenna is actually composed of a series of elementary antennas (linear, circular, etc.) who’s received signals are balanced and combined by using a technique of adaptation in order to control and improve the reception or the transmission. The objective of our study is to look further into the knowledge of the methods of formation of the beams and to elaborate a digital technique of synthesis for the formation of ways and the cancellation of interfering which answers the specifications imposed by the system adaptive by using the networks of neurons. The work of thesis consisted of a contribution to the optimization of the lobe of radiation for an intelligent antenna. In a first stage, two complementary approaches were developed to implement the technique of forming of the lobe, one based on an algorithm of optimization which calculates the ...

This thesis interests in different methods of image compression combining both Bayesian aspects and ``sparse decomposition'' aspects. Two compression methods are in particular investigated. Transform coding, first, is addressed from a transform optimization point of view. The optimization is considered at two levels: in the spatial domain by adapting the support of the transform, and in the transform domain by selecting local bases among finite sets. The study of bases learned with an algorithm from the literature constitutes an introduction to a novel learning algorithm, which encourages the sparsity of the decompositions. Predictive coding is then addressed. Motivated by recent contributions based on sparse decompositions, we propose a novel Bayesian prediction algorithm based on mixtures of sparse decompositions. Finally, these works allowed to underline the interest of structuring the sparsity of the decompositions. For example, a weighting of the decomposition ...

This work deals with the use of an antenna array at the base station of a mobile communication system for transmission. In reception, solutions that exploit the antenna array are now well established. In transmission, however, the problem remains open. Two approaches are possible : exploit the array by using beamforming techniques or by using diversity techniques. These two approaches are based on opposite assumptions about the channels correlation, which implies a greater or smaller distance between antennas, depending on the environment. In practice, these assumptions are not verified. Here, we aim to deal with the problem as a whole for better exploiting the antenna array. This work treats the mono-user case, as well as the multi-user scenario. In the mono-user case, we propose a transmission scheme composed of a classical transmit diversity technique applied to virtual antennas, which are ...

The goal of this work is to propose a general approach for surveillance video indexing and retrieval. Based on the hypothesis that videos are preprocessed by an external video analysis module, this approach is composed of two phases : indexing phase and retrieval phase. In order to profit from the output of various video analysis modules, a general data model consisting of two main concepts, objects and events, is proposed. The indexing phase that aims at preparing data defined in the data model performs three tasks. Firstly, two new key blob detection methods in the object representation task choose for each detected object a set of key blobs associated with a weight. Secondly, the feature extraction task analyzes a number of visual and temporal features on detected objects. Finally, the indexing task computes attributes of the two concepts and stores ...

Visual information, in the form of digital images and videos, has become so omnipresent in computer databases and repositories, that it can no longer be considered a “second class citizen”, eclipsed by textual information. In that scenario, image classification has become a critical task. In particular, the pursuit of automatic identification of complex semantical concepts represented in images, such as scenes or objects, has motivated researchers in areas as diverse as Information Retrieval, Computer Vision, Image Processing and Artificial Intelligence. Nevertheless, in contrast to text documents, whose words carry semantic, images consist of pixels that have no semanticinformation by themselves, making the task very challenging. In this dissertation, we have addressed the problem of representing images based on their visual information. Our aim is content-based concept detection in images and videos, with a novel representation that enriches the Bag-of-Words model. ...

The amount of digital music available both on the Internet and by each listener has considerably raised for about ten years. The organization and the accessibillity of this amount of data demand that additional informations are available, such as artist, album and song names, musical genre, tempo, mood or other symbolic or semantic attributes. Automatic music indexing has thus become a challenging research area. If some tasks are now correctly handled for certain types of music, such as automatic genre classification for stereotypical music, music instrument recoginition on solo performance and tempo extraction, others are more difficult to perform. For example, automatic transcription of polyphonic signals and instrument ensemble recognition are still limited to some particular cases. The goal of our study is not to obain a perfect transcription of the signals and an exact classification of all the instruments ...

To reduce scanning time or improve spatio-temporal resolution in some MRI applications, parallel MRI acquisition techniques with multiple coils have emerged since the early 90’s as powerful methods. In these techniques, MRI images have to be reconstructed from ac- quired undersampled “k-space” data. To this end, several reconstruction techniques have been proposed such as the widely-used SENSitivity Encoding (SENSE) method. However, the reconstructed images generally present artifacts due to the noise corrupting the ob- served data and coil sensitivity profile estimation errors. In this work, we present novel SENSE-based reconstruction methods which proceed with regularization in the complex wavelet domain so as to promote the sparsity of the solution. These methods achieve ac- curate image reconstruction under degraded experimental conditions, in which neither the SENSE method nor standard regularized methods (e.g. Tikhonov) give convincing results. The proposed approaches relies on ...

Retrieval of images that lack a (suitable) annotations cannot be achieved through (traditional) Information Retrieval (IR) techniques. Access through such collections can be achieved through the application of computer vision techniques on the IR problem, which is baptized Content-Based Image Retrieval (CBIR). In contrast with most purely technological approaches, the thesis Human-Centered Content-Based Image Retrieval approaches the problem from a human/user centered perspective. Psychophysical experiments were conducted in which people were asked to categorize colors. The data gathered from these experiments was fed to a Fast Exact Euclidean Distance (FEED) transform (Schouten & Van den Broek, 2004), which enabled the segmentation of color space based on human perception (Van den Broek et al., 2008). This unique color space segementation was exploited for texture analysis and image segmentation, and subsequently for full-featured CBIR. In addition, a unique CBIR-benchmark was developed (Van ...

Image segmentation through deformable models is a method that localizes object boundaries. When difficult segmentation context are proposed because of noise or a lack of information, the use of prior knowledge in the deformation process increases segmentation accuracy. Medical imaging is often concerned by these context. Moreover, medical applications deal with large amounts of data. Then it is mandatory to use a robust and fast processing. This question lead us to a local regularisation of the deformable model. Highly based on the active contour framework, also known as \emph{snake}, we propose a new regularization scheme. This is done by filtering the displacements at each iteration. The filter is based on a smoothing spline kernel whose aim was to approximate a set of points rather than interpolating it. We point out the consistency of the regularization parameter in such a method. ...

Resistivity distribution estimation, widely known as Electrical Impedance Tomography (EIT), is a non linear ill-posed inverse problem. However, the partial derivative equation ruling this experiment yields no analytical solution for arbitrary conductivity distribution. Thus, solving the forward problem requires an approximation. The Finite Element Method (FEM) provides us with a computationally cheap forward model which preserves the non linear image-data relation and also reveals sufficiently accurate for the inversion. Within the Bayesian approach, Markovian priors on the log-conductivity distribution are introduced for regularization. The neighborhood system is directly derived from the FEM triangular mesh structure. We first propose a maximum a posteriori (MAP) estimation with a Huber-Markov prior which favours smooth distributions while preserving locally discontinuous features. The resulting criterion is minimized with the pseudo-conjugate gradient method. Simulation results reveal significant improvements in terms of robustness to noise, computation rapidity ...

In scalable image/video coding, high resolution content is encoded to the highest visual quality and the bit-streams are adapted to cater various communication channels, display devices and usage requirements. These content adaptations, which include quality, resolution and frame rate scaling may also affect the content protection data, such as, watermarks and are considered as a potential watermark attack. In this thesis, research on robust watermarking techniques for scalable coded image and video, are proposed and the improvements in robustness against various content adaptation attacks, such as, JPEG 2000 for image and Motion JPEG 2000, MC-EZBC and H.264/SVC for video, are reported. The spread spectrum domain, particularly wavelet-based image watermarking schemes often provides better robustness to compression attacks due to its multi-resolution decomposition and hence chosen for this work. A comprehensive and comparative analysis of the available wavelet-based watermarking schemes,is performed ...