This paper presents a variational-based approach for fusing hyperspectral and multispectral images. The fusion problem is formulated as an inverse problem whose solution is the target image assumed to live in a lower dimensional subspace. A sparse regularization term is carefully designed, relying on a decomposition of the scene on a set of dictionaries. The dictionary atoms and the supports of the corresponding active coding coefficients are learned from the observed images. Then, conditionally on these dictionaries and supports, the fusion problem is solved via alternating optimization with respect to the target image (using the alternating direction method of multipliers) and the coding coefficients. Simulation results demonstrate the efficiency of the proposed algorithm when compared with state-of-the-art fusion methods.

In this paper, we provide a proof for the conjecture made by Abu Surra et al. [1] to simplify the computation of ensemble input output weight enumerators for protograph-based low density parity check (LDPC) codes. Furthermore, we propose a new method to compute more efficiently the ensemble weight enumerator. This approach can be applied particularly to lighten the computations for high rate codes, generalized LDPC codes or spatially coupled LDPC codes.

Remote sensing images are images of the Earth acquired from planes or satellites. Many different sensors have been developed to image the earth surface, including optical, SAR and hyperspectral images. Change detection on datasets of multitemporal images is one of the main interest of these images. The case where the dataset consist of images acquired by the same sensor has been thoroughly studied, however, dealing with heterogeneous images is very common nowadays. To deal with these images, we proposed a statistical model which describe the joint distribution of their pixel intensity. On unchanged areas, we expect the parameter vector of the model to belong to a manifold. The distance of the model parameter to the manifold can be thus be used as a similarity measure, and the manifold can be learned using images where no changes are present. In this talk I will present the statistical model, its parameter estimation, and the manifold learning approach. The results obtained with this method will be compared with those of other classical similarity measures.

We present a new delay-based transport protocol named FLOWER, that aims at providing a Lower-than-Best-Effort (LBE) service for background traffic using the remaining network capacity. The objective is to propose an alternative to the Low Extra Delay Background Transport (LEDBAT) widely deployed within the official BitTorrent client. Indeed, besides its intra-fairness problem, known as latecomer unfairness, LEDBAT can be too aggressive against TCP, making it ill suited for providing LBE services over certain networks such as constrained wireless networks. By using a fuzzy controller to modulate the sending rate, FLOWER aims to solve LEDBAT issues while fulfilling the role of a LBE protocol. Our simulation results show that FLOWER can carry LBE traffic in network scenarios where LEDBAT can not while solving the latecomer unfairness problem.

End-to-end (E2E) reliable communication in Delay Tolerant Network (DTN) is a challenging task due to long delay and frequent link disruptions. To enable reliability, the IETF is currently looking at strategies to integrate erasure coding mechanisms inside DTN architecture. The objective is to extend the ability of the existing DTN bundle fragmentation mechanism to support cases where bundles have a high probability of being lost. To date, discussions agree that an intermediate node can re-encode bundles, leaving all decoding process at the destination node in order to let intermediate node operations be as simple as possible. We propose to study and analyze possible re-encoding strategies at intermediate nodes using an on-the-fly coding paradigm. We also investigate how re-encoding and acknowledgment strategies based on this coding scheme would enable E2E reliable communication. Finally, we propose an adaptive mechanism with low complexity that deals with both rerouting events and network dynamics which are common in the context of DTN. Simulation results show that re-encoding at the relay and the adaptive mechanism allows a significant reduction in terms of network overhead injected by erasure codes while ensuring the E2E reliability.

The spatial coupling is an efficient technique that improves the threshold of Low Density Parity Check (LDPC) codes. In this paper, we investigate the performance of the serial concatenation of Continuous phase modulation (CPM) and LDPC convolutional codes over a memoryless additive white Gaussian noise channel. We show that coupling protographs optimized for CPM improves their performance and helps designing very good ’small’ protographs. Inspired from convolutional codes and thanks to the inner structure of CPM, we also introduce a new termination without rate loss but that still exhibits a coupling gain and it thus has a very good threshold. We will illustrate the behavior of different LDPC convolutional codes with different termination methods by giving some examples and studying their performance using multidimensional EXIT analysis.

The increasing spatial resolution of hyperspectral remote sensors requires the development of new processing methods capable of combining both spectral and spatial information. In this article, we focus on the spatial component and propose the use of novel multifractal attributes, which extract spatial information in terms of the fluctuations of the local regularity of image amplitudes. The novelty of the proposed approach is twofold. First, unlike previous attempts, we study attributes that efficiently summarize multifractal information in a few parameters. Second, we make use of the most recent developments in multifractal analysis: wavelet leader multifractal formalism, the current theoretical and practical benchmark in multifractalanalysis, and a novel Bayesian estimation procedure for one of the central multifractal parameters. Attributes provided by these stateof-the-art multifractal analysis procedures are studied on two sets of hyperspectral images. The experiments suggest that multifractal analysis can provide relevant spatial/textural attributes which can in turn be employed in tasks such as classification or segmentation.

This paper presents an unsupervised Bayesian algorithm for hyperspectral image unmixing accounting for endmember variability. This variability is obtained by assuming that each pixel is a linear combination of random endmembers weighted by their corresponding abundances. An additive noise is also considered in the proposed model generalizing the normal compositional model. The proposed model is unsupervised since it estimates the abundances and both the mean and the covariance matrix of each endmember. A classification map indicating the class of each pixel is also obtained based on the estimated abundances. Simulations conducted on a real dataset show the potential of the proposed model in terms of unmixing performance for the analysis of hyperspectral images.