Location fingerprinting is a positioning method that exploits the already existing infrastructures such as cellular networks or WLANs. Regarding the recent demand for energy efficient networks and the emergence of issues like green networking, we propose a clustering technique to compress the radio database in the context of cellular fingerprinting systems. The aim of the proposed technique is to reduce the computation cost and transmission load in the mobile-based implementations. The presented method may be called Block-based Weighted Clustering (BWC) technique, which is applied in a concatenated location-radio signal space, and attributes different weight factors to the location and radio components. Computer simulations and real experiments have been conducted to evaluate the performance of our proposed technique in the context of a GSM network. The obtained results confirm the efficiency of the BWC technique, and show that it improves the performance of standard k-means and hierarchical clustering methods.

In this article we propose to illustrate the ability of consensual clustering to provide mining tools in the context of land cover unsupervised classification. The proposed algorithm is based on individual co-association matrices related to several input clusterings that are combined using a Mean Shift optimization procedure. This provides valuable clusters in terms of interpretation and also information about the data to be clustered, which could be useful to discriminate between easily classified pixels and the other ones, requiring human expertise. The interest of our approach is demonstrated using the Boumerdes dataset provided by SERTIT and CNES, in the context of the 2003 earthquake.

The increasing quality of satellite images has generated interests in extracting man-made structures in urban areas, such as buildings and roads. A classification adapted to urban areas can help to identify these structures. In this paper, SAR information are used to improve land-cover classification. We proposed a classification process using both radar and optical data, a segmentation and a classification with Support Vector Machines (SVM).

Semantic coding is used to provide image representation able to make semantics emerge and to learn it. This process is based on a basic decomposition of the image content into ¿visual words¿; using spatial relationships, these visual words are aggregated to form structures and (inter)active learning is used to associate semantics to the emerging structures. A complete OTB tool has been developed to illustrate the whole process; it allows feature extraction and ¿visual words¿ production, as well as LDA and SVM learning approaches.

Estimation of the spectral envelope (magnitude of the transfer function) of a filter driven by a periodic signal is a long-standing problem in speech and audio processing. Recently, there has been a renewed interest in this issue in connection with the rapid developments of processing techniques based on sinusoidal modeling. In this paper, we introduce a new performance criterion for spectral envelope fitting which is based on the statistical analysis of the behavior of the empirical sinusoidal magnitude estimates. We further show that penalization is an efficient approach to control the smoothness of the estimation envelope. In low-noise situations, the proposed method can be approximated by a two-step weighted least-squares procedure which also provides an interesting insight into the limitations of the previously proposed "discrete cepstrum" approach. A systematic simulation study confirms that the proposed methods perform significantly better than existing ones for high pitched and noisy signals.