Earticle

This paper deals with the problem of multispectral image compression. In particular, we propose to substitute the built-in JPEG 2000 wavelet transform by an adequate multiresolution analysis that we devise within the Lifting-Scheme framework. We compare the proposed method to the classical wavelet transform within both multi-2D and full-3D compression strategies. The two strategies are combined with a PCA decorrelation stage to optimize their performance. For a consistent evaluation, we use a framework gathering four families of metrics including the largely used PSNR. Good results have been obtained showing the appropriateness of the proposed approach especially for images with large dimensions.

The tracking of prominent oscillations with time-varying frequencies is a common task in many signal processing applications. Therefore, efficient methods are needed to meet precision and run-time requirements. We propose two methods that extract the energy of the time frequency plane along a cubic spline trajectory. To raise efficiency, a sparse sampling method with dynamically shaped atoms is developed for the one method in contrast to standard Gabor atoms for the other. Numerical experiments using both synthetic and real-life data are carried out to show that dynamic atoms can significantly outperform classical Gabor formulations for this task.

Strain measurement is a quantity used for assessing the regional function of the left ventricular (LV) of the heart. They are computed by tracking the motion of the non-invasive, virtual tags in the cardiac muscle with time. Tracking these tags gives information for each region of the cardiac muscle by quantifying its deformation during contraction (systolic period) and relaxation (diastolic period). However, these strain measurements suffer from inaccuracies caused by the degradation of the tags and the image quality. In this work, numerical simulations are used to investigate the factors contributing to the error in measurements. An empirical model for the estimated strain values is deduced and presented. In addition, a correction method for the measurement errors is proposed based on the empirically-deduced model. The method was validated on real data, and showed potential enhancement by reducing the errors in strain measurements.

In this paper we propose a model of classification based on the principle of the fuzzy proximity of the terms within the documents. Given the heterogeneous nature of the Arabic documents in our possession, we have studied for this purpose the research model based on the semantic proximity of terms and inspired from the classic Boolean model. Our approach is based on the assumption that more the occurrences of terms in query are close with good connectivity in the extracted semantic graph from the set of document , more this document is relevant to this query. We propose a measure that provides a contextual and semantic search. We used not only a semantic graph to highlight the semantic connections between terms, but also an auxiliary dictionary to increase the connectivity of the graph and therefore the discrimination of documents relevant to the query.

Video surveillance applications need tools to track people trajectories. We present here a new method to extract the 3D head trajectory of a person in a room using only one calibrated camera. A 3D ellipsoid representing the head is used to track the person with a hierarchical particle filter. This method can run in quasi-real-time providing reasonable 3D errors for a monocular system.