Sensitivity versus accuracy in ensemble models of Artificial Neural Networks from Multi-objective Evolutionary Algorithms

Neural Computing and Applications - This paper proposes a framework to obtain ensembles of classifiers from a Multi-objective Evolutionary Algorithm (MOEA), improving the restrictions imposed by...     

Quality of dredged material in the river Seine basin (France). II. Micropollutants

Dredging rivers is needed to ensure safe navigable waters, rivers and waterways. To anticipate the management of dredged materials in the case of the river Seine basin, the quality of the sediments in the river is checked every 3 years before dredging operations. The river Seine Basin is heavily submitted to pollution pressure from nearby industrial activities and urban expansion of Paris and its region. Here, the micropollutant content of the sediment sampled in 1996, 1999 and 2000 before dredging is discussed compared to regulatory standards. The results indicate that most of the sediment samples from the river Seine basin are lightly to moderately contaminated with organic and inorganic micropollutants (heavy metals, PAH, PCB), which makes the management after dredging easier. This pollution is strongly correlated with the organic matter content and to the fine fraction (<50 microm) of the sediment. These results can lead to other management options than the ones already used in the river Seine basin: (1) dumping of lightly to moderately polluted sediments in quarries; and (2) physical treatment (sieving, hydrocycloning) of contaminated sediments issued from 'hot spots'.     

Strain and Stress Fields During Scratch Tests on Amorphous Polymers: Influence of the Local Friction

In this paper, we present results deduced from three-dimensional finite element simulations of scratching, with spherical indenter geometry at different imposed ratios, a/R in the range of 0.1–0.9. For each simulated ratio a/R, the local friction has been increased from 0 to 1. The paper aims at studying the tangential scratch behaviour of homogeneous polymeric substrates, considered in first approximation as elastic linear-hardening plastic material. For only elastic–plastic contacts, without any strain rate or temperature effects, it focuses on studying some characteristic response due to spherical scratching process as a function of scratching conditions (a/R, μ _loc ) such as the stress and plastic strain fields, including the plastic zone dimension and the definition of an volume average plastic strain.     

A 65nm CMOS Ramp Generator Design and its Application Towards a BIST Implementation of the Reduced-Code Static Linearity Test Technique for Pipeline ADCs

This work presents an efficient on-chip ramp generator targeting to facilitate the deployment of Built-In Self-Test (BIST) techniques for ADC static linearity characterization. The proposed ramp generator is based on a fully-differential switched-capacitor integrator that is conveniently modified to produce a very small integration gain, such that the ramp step size is a small fraction of the LSB of the target ADC. The proposed ramp generator is employed in a servo-loop configuration to implement a BIST version of the reduced-code linearity test technique for pipeline ADCs, which drastically reduces the volume of test data and, thereby, the test time, as compared to the standard test based on a histogram. The demonstration of the pipeline ADC BIST is carried out based on a mixture of transistor-level and behavioral-level simulations that employ actual production test data.     

Comparison of applied and induced current electrical impedance tomography

Several papers on induced current electrical impedance tomography (IC-EIT) have dwelt on potential advantages of this technique over conventional EIT which uses applied current (AC-EIT). Experimental evidence that IC-EIT could surpass AC-EIT in similar imaging conditions is lacking. In this paper, we describe a system that can switch rapidly between both AC-EIT and IC-EIT. The system makes it possible to image objects in a saline-filled tank, providing data acquired in identical test conditions for comparing the performance of the two modes. The system uses eight circular coils and 16 electrodes to acquire 120 linearly independent measurements in IC-EIT and 104 in AC-EIT. Difference images were reconstructed from data acquired with both modes using the maximum a posteriori method. Spatial resolution was lower in IC-EIT images than in AC-EIT, especially in the radial direction. IC-EIT also exhibits a bias toward the center for positioning a conductivity perturbation. These results were obtained for a typical coil configuration widely used in the literature and may not be representative of alternate coil configurations. The system described in this paper provides stable experimental conditions for comparing the performance of the two EIT imaging modes and would be a valuable tool for validating new coil configurations.     

Retrieval of pigment concentrations and size structure of algal populations from their absorption spectra using multilayered perceptrons

Spectral absorption coefficients of phytoplankton can now be derived, under some assumptions, from hyperspectral ocean color measurements and thus become accessible from space. In this study, multilayer perceptrons have been developed to retrieve information on the pigment composition and size structure of phytoplankton from these absorption spectra. The retrieved variables are the main pigment groups (chlorophylls a, b, c, and photosynthetic and nonphotosynthetic carotenoids) and the relative contributions of three algal size classes (pico-, nano-, and microphytoplankton) to total chlorophyll a. The networks have been trained, tested, and validated using more than 3,700 simultaneous absorption and pigment measurements collected in the world ocean. Among pigment groups, chlorophyll a is the most accurately retrieved (average relative errors of 17% and 16% for the test and validation data subsets, respectively), while the poorest performances are found for chlorophyll b (average relative errors of 51% and 40%). Relative contributions of algal size classes to total chlorophyll a are retrieved with average relative errors of 19% to 33% for the test subset and of 18% to 47% for the validation subset. The performances obtained for the validation data, showing no strong degradation with respect to test data, suggest that these neural networks might be operated with similar performances for a large variety of marine areas.     

General Object Reconstruction Based on Simplex Meshes

In this paper, we propose a general tridimensional reconstruction algorithm of range and volumetric images, based on deformable simplex meshes. Simplex meshes are topologically dual of triangulations and have the advantage of permitting smooth deformations in a simple and efficient manner. Our reconstruction algorithm can handle surfaces without any restriction on their shape or topology. The different tasks performed during the reconstruction include the segmentation of given objects in the scene, the extrapolation of missing data, and the control of smoothness, density, and geometric quality of the reconstructed meshes. The reconstruction takes place in two stages. First, the initialization stage creates a simplex mesh in the vicinity of the data model either manually or using an automatic procedure. Then, after a few iterations, the mesh topology can be modified by creating holes or by increasing its genus. Finally, an iterative refinement algorithm decreases the distance of the mesh from the data while preserving high geometric and topological quality. Several reconstruction examples are provided with quantitative and qualitative results.     

Refractive effects in coherent anti-Stokes Raman scattering microscopy

Coherent anti-Stokes Raman scattering (CARS) microscopy with high sensitivity and high three- dimensional resolution has been developed for the vibrational imaging of chemical species. Due to the coherent nature of the CARS emission, it has been reported that the detection of epi-CARS and forward-CARS (F-CARS) signals depends on the size and shape of the sample. We investigate theoretically and experimentally the effects on the CARS signal of refractive index mismatches between the sample and its surroundings. Backward-CARS and F-CARS signals are measured for different polystyrene bead diameters embedded in different refractive index solvents. We show that index mismatches result in a backward-reflected F-CARS signal that generally dominates the experimentally backward-detected signal. Simulations based on geometrical and wave optics comparing forward- and backward-detected signals for polystyrene beads embedded in different index solvents confirm our findings. Furthermore, we demonstrate that the maxima of forward- and backward-detected signals are generated at different positions along the optical axis in the sample if refractive index mismatches are present between the sample and its surroundings.     

SOCRATE: an optical bench dedicated to the understanding and improvement of a laser conditioning process

We present an automatic excimer laser bench (SOCRATE) allowing for the treatment of optical components by laser conditioning. This apparatus, developed at the Commissariat a l'Energie Atomique-Le Ripault, has been designed to add to this conditioning process an in situ, accurate laser-induced damage threshold (LIDT) measurement and different nondestructive optical techniques for the characterization of the component during treatment. Through different examples, we demonstrate the importance of these characterizations to improve the understanding of the laser conditioning. The role of an in situ adapted metrology associated in real time with a laser conditioning bench offers new opportunities to analyze laser-induced damage mechanisms and subsequently to increase the LIDT of optical components.