PASEO Group / People / Christophe Collet

FILTER: Display only PASEO group-related publications

Show link boxes

Journal papers and Book Chapters – peer-reviewed

C. Collet, F. Flitti, S. Bricq, A. Jalobeanu: “Fusion and Multi-Modality” - in Inverse Problems in Vision and 3D Tomography (ISTE/Wiley), A. Mohammad-Djafari ed., John Wiley and Sons, Dec 2009
@inbook{ref86,
}
F. Salzenstein, C. Collet, S. Le Cam, M. Hatt: “Non Stationary Fuzzy Markov Chains” - Pattern Recognition Letters (PRL), in press, Jul 2007
This paper deals with a recent statistical model based on fuzzy Markov random chains for image segmentation, in the context of stationary and non-stationary data. On one hand, fuzzy scheme takes into account discrete and continuous classes through the modeling of hidden data imprecision and on the other hand, Markovian Bayesian scheme models the uncertainty on the observed data. A non-stationary fuzzy Markov chain model is proposed in an unsupervised way, based on a recent Markov triplet approach. The method is compared with the stationary fuzzy Markovian chain model. Both stationary and non-stationary methods are enriched with a parameterized joint density, which governs the attractiveness of the neighbored states. Segmentation task is processed with Bayesian tools, such as the well known MPM (Mode of Posterior Marginals) criterion. To validate both models, we perform and compare the segmentation on synthetic images and raw optical patterns which present diffuse structures.
@article{ref72,
}
F. Salzenstein, C. Collet: “Fuzzy Markov Random Fields versus Chains for Multispectral Image Segmentation” - IEEE Trans. on PAMI (TPAMI), 28(11), Nov 2006
This paper deals with a comparison of recent statistical models based on fuzzy Markov random fields and chains for multispectral image segmentation. Fuzzy scheme takes into account discrete and continuous classes which models the imprecision of the hidden data. In this framework, we assume the dependence between bands and we express the general model for the covariance matrix. A fuzzy Markov chain model is developed in an unsupervised way. This method is compared with the fuzzy Markovian field model previously proposed by one of the authors. Segmentation task is processed with Bayesian tools, such as the well known MPM (Mode of Posterior Marginals) criterion. Our goal is to compare the robustness and rapidity for both methods (fuzzy Markov fields versus fuzzy Markov chains). Indeed, such fuzzy-based procedures seem to be a good answer, e.g., for astronomical observations when the patterns present diffuse structures. Moreover, these approaches allow us to process missing data in one or several spectral bands which correspond to specific situations in Astronomy. To validate both models, we perform and compare the segmentation on synthetic images and raw multispectral astronomical data.
@article{ref49,
}
F. Flitti, C. Collet, B. Vollmer, F. Bonnarel: “Multiband segmentation of a spectroscopic line data cube: application to the HI data cube of the spiral galaxy NGC 4254” - EURASIP journal on Applied Signal Processing (JASP), special issue on Applications of Signal Processing in Astrophysics and Cosmology, 2005(15), Jun 2005
A new method for the multiband segmentation of a spectroscopic line data cube is presented. This method is intended to help astronomers to handle complex spectroscopic line data cubes where the inspection of the channel and moment maps is difficult. Due to the Hughes phenomenon, the number of input images for the segmentation process is limited. Therefore, the spectrum of each pixel is fitted with a mixture of 6 Gaussians with fixed mean values and variances. The maps of the Gaussian weights are the input for a Markovian segmentation algorithm. The final segmentation map contains classes of pixels with similar spectral line profiles. The application of our method to the HI data cube of the Virgo spiral galaxy NGC 4254 shows that kinematically interesting regions can be detected and masked by our method.
@article{ref14,
}
M. Petremand, M. Louys, C. Collet: “Color display for multiwavelength astronomical images” - Traitement du signal (TS), Numéro spécial, 21(6), Jan 2005
This paper proposes a new approach for the color display of multispectral/ hyperspectral images. The color representation of such data becomes problematic when the number of bands is higher than three, i.e. the basic RGB (Red, Green, Blue) representation is not straightforward. Here we employ a technique that uses a segmentation map, like an a priori information, and then compute a Factorial Discriminant Analysis (Fischer analysis) in order to allow, at best, a distribution of the information in the color space HSV (Hue, Saturation, Value). The information collected from the segmentation map (where each pixel is associated with class) has been shown to be advantages in the representation of the images through the results obtained on increasing size image collections in the framework of astronomical images. This method can easily be applied to other domains such as polarimetric or remote sensing imagery.
@article{ref15,
}
C. Collet, F. Murtagh: “Segmentation based on a hierarchical Markov model” - Pattern Recognition (PR), 37(12), Dec 2004
We develop a new multiscale Markov segmentation model for multiband images. Using quadtree multiple resolution analysis of a multiband image, we use both inter- and intra-scale spatial Markov statistical dependencies. Bayesian inference is used to assess the appropriate number of segments. We exemplify the excellent results which can be obtained with this approach using synthetic images, and in two case studies involving multiband astronomical image sets.
@article{ref19,
}
C. Collet, Y. Takakura, J. Zallat: “Clustering of Mueller matrix images for skeletonized structure detection” - Optics Express (OE), 12(7), Apr 2004
This paper extends and refines previous work on clustering of polarization-encoded images. The polarization-encoded images used in this work are considered as multidimensional parametric images where a clustering scheme based on Markovian Bayesian inference is applied. Hidden Markov Chains Model (HMCM) and Hidden Hierarchical Markovian Model (HHMM) show to handle effectively Mueller images and give very good results for biological tissues (vegetal leaves). Pretreatments attempting to reduce the image dimensionality based on the Principal Component Analysis (PCA) turns out to be useless for Mueller matrix images.
@article{ref16,
}
F. Salzenstein, C. Collet, M. Petremand: “Fuzzy Markov Random Fields for multispectral images” - Traitement du signal (TS), 21(1), Jan 2004
This paper deals with a new statistical segmentation based on fuzzy multispectral markovian random fields model.We propose to solve the problem of parameter estimation, applying a stochastic gradient algorithm and empirical moment method, in order to estimate respectively the a priori parameters of the hidden Markovian field and the conditional densities of the observed data. Under correlated spectral band assumption, we introduce a model to express the variance-covariance matrix related to the fuzzy classes, by means of the ones related to the hard classes. We compare the results applying MPM (Mode of Posterior Marginales) and ICM (Iterated Conditional Mode) algorithms. We validate our procedure on synthetic images and test this approach on real multispectral astronomical data.
@article{ref17,
}
J. Zallat, C. Collet, Y. Takakura: “Clustering of Polarization-Encoded Images” - Applied Optics (AO), 43(2), Jan 2004
Polarization-encoded imaging consists of the distributed measurements of polarization parameters for each pixel of an image. We address clustering of multidimensional polarization-encoded images. The spatial coherence of polarization information is considered. Two methods of analysis are proposed: polarization contrast enhancement and a more-sophisticated image-processing algorithm based on a Markovian model. The proposed algorithms are applied and validated with two different Mueller images acquired by a fully polarimetric imaging system.
@article{ref18,
}

Conference papers – peer-reviewed

M. Louys, B. Perret, B. Vollmer, F. Bonnarel, S. Lefèvre, C. Collet: “First results on a dedicated extraction pipeline based on bayesian segmentation and luminosity analysis profiles” - Astronomical Data Analysis Software & Systems XVII (ADASS'07), London, UK, Sep 2007
We have designed a new technique for the detection of Low Surface Brightness galaxies based on local background/source separation using Markovian analysis. This method helps to estimate smooth local variations of the background and therefore allows for determining source candidates as faint as LSB galaxies. For each source an average density profile is computed, the shape of which can help to sort out stars and bright ob jects. A list of LSB candidates is provided, for which position, profile and surface brightness are examined thoroughly. The results are very promising. This approach has been compared to the SExtractor source detection tool and to a previous original analysis by S.Sabatini et al. on the same INT image dataset of the Virgo Cluster. Detection rate, source selection criteria and calculation loop improvements are discussed.
@inproceedings{ref78,
}
E. Aptoula, S. Lefèvre, C. Collet: “Mathematical Morphology applied to the segmentation and classification of galaxies in multispectral images” - European Signal Processing Conference (EUSIPCO'06), Florence, Italy, Sep 2006
The automated segmentation and classification of galaxies still constitute open problems for astronomical imaging, mainly due to their fuzzy and versatile nature, as well as to the multitude of the available channels. In this paper, a mathematical morphology based approach is explored. First, a semi-automated method for multispectral galaxy segmentation, based on the marker controlled watershed transformation is proposed. Moreover, a novel and viewpoint independent morphological feature, based on the top-hat operator, is introduced for the distinction of spiral from elliptical galaxies. Illustrative application examples of the presented approach on actual images are also provided.
@inproceedings{ref50,
}
F. Flitti, C. Collet: “Markov Regularization of Mixture of Latent variable Models for Multi-component Image Unsupervised Joint Reduction/Segmentation” - 9th International Conference on Information Fusion (FUSION'06), Florence, Italy, Jul 2006
This paper is concerned with Multicomponent image segmentation which plays an important role in many imagery applications. Unfortunately, we are faced with Hughes phenomenon when the number of components increases, and one often carry out a space dimensionality reduction as a preprocessing step before segmentation. An interesting solution is the mixtures of latent variable models which recover clusters in the observation structure and establish a local linear mapping on a reduced dimension space for each cluster. Thus a globally nonlinear model is obtained to reduce dimensionality. Furthermore, a likelihood to each local model is often available which allows a well formulation of the mixture model and a maximum likelihood based decision for the clustering task. However for D-component images classification, such clustering, based only on the distance between observations in the D-dimensional space is not adapted since it neglects the observation spatial locations in the image. We propose to use a Markov a priori associated with such models to regularize D-dimensional pixel classification. Thus segmentation and reduction are performed simultaneously. In this paper, we focus on the Probabilistic Principal Component Analysis (PPCA) as latent model, and the Hidden Markov quad-Tree (HMT) as a Markov a priori.
@inproceedings{ref51,
}
F. Flitti, C. Collet, A. Joannic-Chardin : “Unsupervised Multiband Image Segmentation using Hidden Markov Quadtree and Copulas” - IEEE International Conference on Image Processing (ICIP'05), Genova, Italy, Sep 2005
This paper deals with Hidden Markov Quadtree model for multiband image segmentation. This task, requiring multivariate probability density computations for the data likelihood term, is often confronted with the lack of analytical multidimensional expressions in the non-gaussian case. Thus, multidimensional Gaussian distribution is usually used for its simplicity, even if Gaussian assumption is not always verified. In this work, we propose a new approach based on copula theory to compute multivariate density on Markov Quadtree.
@inproceedings{ref21,
}
F. Flitti, C. Collet, E. Slezak: “Wavelet domain astronomical multiband image fusion and restoration using Markov quadtree and copulas” - 13th European Signal Processing Conference (EUSIPCO'05), Antalya, Turkey, Sep 2005
Fusion of multiband images is of great interest in several applications like astronomy, remote sensing and medecine. It allows to obtain an efficient summary of the whole multiband information in a single scene. Obviously, fusion is more difficult for noisy observations. On one hand multiscale analysis is a popular choice in recent fusion research. On the other hand, wavelet framework is very well adapted for denoising task. Thus wavelet domain seems a quite appropriate for noisy image fusion. Recently, an efficient wavelet Markov modeling was introduced, capturing interscale and spatial wavelet coefficient correlations. In this paper we use a more general Markovian framework, modeling not only spatial and interscale dependencies as the existent models do but also interband correlation for multiband image joint fusion and denoising. Moreover, the multidimensional likelihood is modeled using the copulas theory which allows us to use any kind of marginal densities with a given interband correlation.
@inproceedings{ref22,
}
F. Flitti, C. Collet, E. Slezak: “Astronomical multiband image fusion and restoration using pyramidal analysis and markovian segmentation” - 8th International Conference on Information Fusion (ISIF'05), Philadelphia, PA, USA, Jul 2005
This paper deals with noised astronomical multiband image fusion and restoration. The wavelet domain is well adapted for such tasks. In fact, intensity fluctuations corresponding to the noise are most important at the finest resolution and related details coefficients decrease quickly as the scale increases. Real structures in the image will therefore lead to larger detail coefficients values. This information can be suitably combined to detect and fuse real structures in all bands. In this work, we present a new fusion-restoration scheme for astronomical multispectral data, operating in four steps. At first, a pyramidal algorithm with one wavelet analyzes all spectral bands leading to a pyramid of details coefficients for each one. Secondly, the detail information for all wavelengths are combined in a unique Multiband-Multiresolution Pyramidal to feed a hierarchical Markovian classifier based on a quadtree topology, providing a multiscale two classes segmentation map masking at each scale small enough coefficients. At third, the selected coefficients are merged leading to a unique pyramidal structure of detail coefficients. Finally, this resulting structure is used with an iterative reconstruction procedure to recover the denoised fused image. The Markovian quadtree models spatial, interscale and interband correlation together allowing a real whole cube information fusion.
@inproceedings{ref20,
}
C. Collet, F. Flitti: “Variations on Markovian Quadtree Model for Multiband Astronomical Image Analysis” - Int. Symp. on applied stochastic models and data analysis (ASMDA'05), Brest - Le Quartz, May 2005
This paper is concerned with the analysis of multispectral observations, provided by space or ground telescopes. The large amount and the complexity of heterogeneous data to analyse lead us to develop new methods for segmentation tasks, which aim to be robust, fast and e±cient. Some prior knowledge on the information to be extracted from the original image is available, and Bayesian statistical theory is known to be a convenient tool to take this a priori knowledge into consideration. In this paper, we investigate the use of the Bayesian inference on Markovian quadtrees for some reduction, fusion, segmentation or restoration problems of great importance in multiband astronomical imagery.
@inproceedings{ref23,
}
F. Flitti, C. Collet, B. Vollmer, F. Bonnarel: “Data reduction of hyperspectral radio-astronomical images for galaxy cluster segmentation” - 4th Int. Conf. on Physics in Signal and Image Processing (PSIP'05), Toulouse, France, Feb 2005
This paper proposes a reduction-segmentation scheme for radioastronomical cubes. In order to avoid the curse of dimensionality phenomenon, a reduction technique is proposed as preprocessing step before classification. On each site of the image a spectrum is observed, exhibiting few spectral rays, modeled as a weighted mixture of selected Gaussian functions. These weights feed a Hierarchical Markovian classifier, in order to cluster spatially homogeneous areas with similar spectra behaviors. Such approach is very useful in radioastronomical context, because it allows to highlight regions of astronomical interest where astrophysical investigations may focus.
@inproceedings{ref24,
}
C. Collet, M. Louys, A. Oberto, C. Bot: “Markov Model for Multispectral Image analysis: application to Small Magellanic Cloud segmentation” - International Conference on Image Processing (ICIP'03), Barcelona, Spain, Sep 2003
This paper deals with the unsupervised segmentation of astronomical multiband images. Most of these images have the particularity to be quantized on float numbers with large luminance range, on different wavelengths. These characteristics require to manipulate large amount of extremely accurate data on each spectral band, which is very different to the case of 8-bits-integer coded pixels. We present some results obtained on Multiwavelength images of the Small Magellanic Cloud, by using the Marginal Posterior Mode (MPM) estimator on a quadtree structure under Markovian assumption. The estimation of the model parameters is then addressed with Expectation-Maximization (EM)-type algorithms, allowing unsupervised hyperparameter estimation. The main interest of this modeling effort lies in its generality : the algorithm handles multiwavelength floating data in a single upward and downward scan on the quadtree. A new aspect in this paper concerns the noise statistics that are supposed to be lognormal for each class. Another new aspect, is the in-scale-coding of the label map.
@inproceedings{ref27,
}
F. Flitti, C. Collet: “ACP et ACI pour la réduction de données en imagerie astronomique multispectrale” - 19ème colloque sur le traitement du signal et des images (GRETSI'03), Paris, France, Sep 2003
A technique of astronomical data reduction as a preliminary stage of farway galaxies images classification is presented. The retained classifier is based on a in-scale causal Markovian modeling on the quadtree. This paper presents the encouraging results obtained by this approach.
@inproceedings{ref29,
}
M. Louys, C. Bot, A. Oberto, C. Collet: “Multiwavelength image analysis of the Small Magellanic Cloud using hierarchical Markovian segmentation” - 3rd workshop on Physics in Signal and Image Processing (PSIP'03), Grenoble, France, Jan 2003
This paper addresses the segmentation of astronomical multiband images with missing data. We present some results obtained on multiwavelength images of the Small Magellanic Cloud, by using the Marginal Posterior Mode (MPM) estimator on a quadtree structure under Markovian assumption : the estimation of the model parameters is then addressed with Expectation-Maximization (EM)-type algorithms, allowing unsupervised hyperparameter estimation. The main interest of this modeling effort lies in its generality: the algorithm handles multiwavelength data (possibly with missing data) in a single Causal-in-scale Markovian model. It is an interesting tool for astronomical image analysis, which exhibits very large dynamic range of intensities and missing data on the sampling grid in this case.
@inproceedings{ref30,
}
F. Murtagh, C. Collet, M. Louys, J-L. Starck: “Multiresolution Filtering and Segmentation of Multispectral Images” - Proc. of SPIE, Astronomical Data Analysis III (SPIE-4847), Waikoloa, Hawaii, USA, Aug 2002
@inproceedings{ref32,
}
C. Collet, M. Louys, J-N. Provost, A. Oberto: “Fusion of Astronomical Multiband Images on a Markovian Quadtree” - Information Fusion (Fusion'02), Annapolis, Maryland, USA, Jul 2002
Over the last few years, an enormous amount of studies has been devoted to astronomical image restoration. The new challenge, nowadays, has been reported on the use of multichannel techniques to restore, segment and classify such images. The multichannel image segmentation problem, is a new field of great interest for the astronomical community. This paper presents an unsupervised method to automatically segment multispectral astronomical images, which is the first stage toward the automatic classification of astronomical objects. The proposed method relies on a hierarchical Markovian modeling on a quadtree, including the estimation of all involved parameters. The parameters of the prior model are automatically calibrated while the estimation of the noise parameters is solved by identifying generalized distribution mixtures, using an Iterative Conditional Estimation procedure. Generalized Gaussian distributions are considered to model intensity distribution of the multispectral images.
@inproceedings{ref33,
}

Abstracts, Posters, Preprints, Reports and Theses

C. Collet: “Contribution à l'analyse statistique des images en acoustique sous-marine et en océanographie” - HDR (research advisor qualification), Univ. of Bretagne Occidentale, France, May 2000
à compléter
@unpublished{ref13,
}

Who Am I?

Educational Background

Academic Positions

Research Interests

Applications I am interested in

Favorite Approaches & Methods

PhD theses I am currently advising

Related Research Projects

Astronomy & Astrophysics

Planetary Imaging (Small Bodies & Planets)

Remote Sensing (Earth Observation)

Related Software Projects

Selected Publications

Journal papers and Book Chapters – peer-reviewed

Conference papers – peer-reviewed

Abstracts, Posters, Preprints, Reports and Theses

Links