Query by visual example (QBVE) has been widely exploited in image retrieval. If starting image is missing, the query by visual thesaurus paradigm allows the user to compose his mental query image through visual patches summarizing the region database. Researches on the human visual system have provided considerable evidence that the color and texture must be processed separately. In this paper, we propose to enrich the paradigm of mental image search by constructing a hierarchical visual thesaurus of the regions provided by a new region labeling criterion into homogeneous and textured regions for boosting the object recognition. The new labeling criterion is based on the spatial dispersion of interest points in the region. Our point based criterion has been validated on a satellite image database. We can prove that our approach is able to retrieve complex concepts better than describing homogeneous and textured regions with the same visual feature.

In this work we demonstrate the potentiel of the InSAR coherence in classifying regions threatened by the desertification. We propose to combine features from InSAR amplitude and coherence images into a single 3D space in which every pixel is represented by its three-values vector (master image, slave image, coherence image) then to apply the multichannel Fuzzy C-Means (FCM) clustering algorithm. Our motivation behind this proposition lies first in the fact that correlation between both of InSAR images can reduce considerably effects of the noise while preserving image details. The second fact is that coherence represents a complementary information to the SAR image amplitude that can well indicates the spatial homogeneity behavior of each contextual class. Two ERS-1 SAR images over the region of Ben Guerdane, in the South-East of Tunisia, are considered for validation.

In this paper, we present a probabilistic framework for edge and region grouping using conditional random field. Our model is built on a hybrid adjacency graph of atomic region and contour primitives. Unary and pairwise potentials that capture similarity, proximity and curvilinear continuity are defined. Similarity, for both region and edge cues, is measured by likelihood ratios learned from a human labeled ground truth. We use a stochastic graph partition algorithm, Swendsen-Wang Cut, to perform inference on this model. Experimental results are shown on gray-scale natural images.

In this paper, we propose a new Bayesian estimation-based algorithm for two dimensional phase unwrapping of discontinuous phase fields with noisy principal values. The proposed algorithm uses the Markov random field models to build the prior distribution, so the unwrapping problem became equivalent to a minimization of an energy function. Our main contribution in this work is to propose a modification to the classical quadratic potential function, which enforces a global smoothness condition, so that the phase jumps, which result from the phase discontinuities, contribution to the energy are mitigated. This was possible throw weighting the classical quadratic potential function by the probability of occurrence of these jumps which decrease when they increase. Theoretically, this probability follows an hypergeometric distribution which can be approximated as a Gaussian one in order to make easier mathematical manipulations. An analytical expression for the minimization automate was derived.

We propose a non-homogeneous conditional random field (CRF) built over an adjacency graph of superpixels for contextual region grouping. Our model includes spatially dependent potentials that capture contextual interactions of the data as well as the labels. Both superpixels and segments are described with local statistics which take into account their contexts in the image. This results the non-homogeneity of the fields which improves the region grouping process of natural images. In our energy formulation, the similarity is measured by a likelihood ratio learned from a human labeled ground truth. The inference is performed using a cluster sampling method, the Swendsen-Wang cut algorithm. Results are shown on various natural images.

We address unsupervised variational segmentation of multi-look complex polarimetric images using a Wishart observation model via level sets. The methods consists of minimizing a functional containing an original data term derived from maximum likelihood Wishart approximation and a classical boundary length prior. The minimization is carried out efficiently by first order expansion of the data term and a new multiphase method which embeds a simple partition constraint directly in curve evolution. Results are shown on both synthetic and real images. Quantitative performance evaluation and comparisons with another method are also given.

This study investigates a level set method for complex polarimetric image segmentation. It consists of minimizing a functional containing an original observation term derived from maximum-likelihood approximation and a complex Wishart/Gaussian image representation and a classical boundary length prior. The minimization is carried out efficiently by a new multiphase method which embeds a simple partition constraint directly in curve evolution to guarantee a partition of the image domain from an arbitrary initial partition. Results are shown on both synthetic and real images. Quantitative performance evaluation and comparisons are also given.

We investigate a method of segmentation of multichannel polarimetric images, such as in laser illuminated or synthetic aperture radar, into a given but arbitrary number of regions via curve evolution and level sets. The algorithm consists of evolving closed curve, within an explicit correspondence between the interiors of curves and regions segmentation, to minimize a multivariate criterion corresponding to the complex Gaussian polarimetric model and a term of smoothness of the boundaries of regions. Results are shown on a polarimetric image.

The purpose of this study is to investigate Synthetic Aperture Radar (SAR) image segmentation into a given but arbitrary number of gamma homogeneous regions via active contours and level sets. The segmentation of SAR images is a difficult problem due to the presence of speckle which can be modeled as strong, multiplicative noise. The proposed algorithm consists of evolving simple closed planar curves within an explicit correspondence between the interiors of curves and regions of segmentation to minimize a criterion containing a term of conformity of data to a speckle model of noise and a term of regularization. Results are shown on both synthetic and real images.

Motion-based segmentation is an important task in object-oriented video applications. Different approaches to motion segmentation with level sets. have been proposed. The key features of level sets representation are its ability to handle variations in the topology of the segmentation and its numerical stability. These approaches rely on certain prior information to perform motion segmentation. We present a new algorithm for segmenting an image into distinct regions of homogeneous motion. The two problems of motion estimation and motion segmentation are jointly solved without need of prior information. We provide experimental results on image sequences with synthetic and natural motion.

The aim of this present work is to find an inverse model to retrieve roughness geometric and dielectric parameters of natural rough surfaces from radar backscattering data. The bi-dimensional surfaces are described by means of the fractional Brownian motion random process, using the bi-dimensional wavelet transform. Multi-scale roughness is characterized by two parameters, the first one proportional to the standard deviation and the other one related to the fractal dimension. Soil moisture is related to the complex dielectric constant. To simulate radar backscattering we used the small perturbation model in which the radar backscattering coefficient can be expressed as the product of two factors, the first dependent on polarisation but independent of surface roughness and vice versa the second dependent on roughness but independent of polarisation. Thus, this model simplifies the procedure of inversion and the co-polarised ratio between hh and vv polarisation is independent of roughness and a minimisation over only two parameters is performed to retrieve the complex dielectric constant independently of the employed geometric surface description. Once the dielectric constant is known, the retrieval of multi-scale surface roughness parameters is performed in a successive step by using multi-frequency and multi-incident angle data.

We study two unsupervised algorithms for polarimetric SAR image classification. The first one is Cloude's decomposition algorithm. The main advantage of this unsupervised algorithm is to provide terrain identification information where the most important kinds of scattering medium can be discriminated. However, his main advantage is the arbitrary location of decision boundaries. To surmount this insufficiency, we present the second algorithm based on neural networks. We propose a new scheme of unsupervised classification that combine the most important kind of trained nets.

The principal objective of the present work is to find an adequate description of natural rough surfaces in order to study backscattering from these surfaces and try in a future work to retrieve soil characteristic parameters. Previous work has shown the inadequacy of the classical description which considered rough surfaces as stationary random Gaussian processes. They proposed a multi-scale surface description considering one-dimensional surfaces and incorporated it into the integral equation model (IEM). We extend this multi-scale description to the bi-dimensional case using the Mallat algorithm and we investigate its impact on radar backscattering.

This paper describes the use of polarimetric data for Earth terrain classification. The Cloude (1992) target decomposition theorem is applied to classify the scene into several classes having each one a distinct physical interpretation. The implemented algorithm, under its new version, uses two parameters, not often exploited in radar remote sensing, namely the entropy and a parameter linked to scattering mechanisms. The used data are polarimetric and multifrequency and are extracted from the Landes forest in France. The obtained results are satisfactory since one can understand the physical behavior of the wave when it is in contact with an obstacle.

This paper describes the use of polarimetric SAR data for Earth terrain identification and classification. The K-distribution is used to describe the spatial characteristics of a forest area. Statistical methods using supervised and unsupervised techniques are also applied to segment the simple and the polarimetric data. The supervised classification is based on the ML algorithm assuming a Rayleigh distribution for the intensity pixels of the HH image and a circular Gaussian for the full polarimetric data. For the unsupervised classification, we use both the clustering and the ICM (iterative conditional modes) algorithms. Using such techniques, the segmentation we could classify areas of trees which differ by their age and which have in general an approachable scattering behaviour.>