Parallel hyperspectral unmixing method via split augmented lagrangian on GPU

One of the main problems of hyperspectral data analysis is the presence of mixed pixels due to the low spatial resolution of such images. Linear spectral unmixing aims at inferring pure spectral signatures and their fractions at each pixel of the scene. The huge data volumes acquired by hyperspectral sensors put stringent requirements on processing and unmixing methods. This letter proposes an efficient implementation of the method called simplex identification via split augmented Lagrangian (SISAL) which exploits the graphics processing unit (GPU) architecture at low level using Compute Unified Device Architecture. SISAL aims to identify the endmembers of a scene, i.e., is able to unmix hyperspectral data sets in which the pure pixel assumption is violated. The proposed implementation is performed in a pixel-by-pixel fashion using coalesced accesses to memory and exploiting shared memory to store temporary data. Furthermore, the kernels have been optimized to minimize the threads divergence, therefore achieving high GPU occupancy. The experimental results obtained for the simulated and real hyperspectral data sets reveal speedups up to 49 times, which demonstrates that the GPU implementation can significantly accelerate the method's execution over big data sets while maintaining the methods accuracy.

A DC-DC Step-Up mu-Power Converter for Energy Harvesting Applications, Using Maximum Power PointTracking, Based on Fractional Open Circuit Voltage

A DC-DC step-up micro power converter for solar energy harvesting applications is presented. The circuit is based on a switched-capacitorvoltage tripler architecture with MOSFET capacitors, which results in an, area approximately eight times smaller than using MiM capacitors for the 0.131mu m CMOS technology. In order to compensate for the loss of efficiency, due to the larger parasitic capacitances, a charge reutilization scheme is employed. The circuit is self-clocked, using a phase controller designed specifically to work with an amorphous silicon solar cell, in order to obtain themaximum available power from the cell. This will be done by tracking its maximum power point (MPPT) using the fractional open circuit voltage method. Electrical simulations of the circuit, together with an equivalent electrical model of an amorphous silicon solar cell, show that the circuit can deliver apower of 1132 mu W to the load, corresponding to a maximum efficiency of 66.81%.

FMRI 3D registration based on Fourier space subsets using neural networks

In this work, we present a neural network (NN) based method designed for 3D rigid-body registration of FMRI time series, which relies on a limited number of Fourier coefficients of the images to be aligned. These coefficients, which are comprised in a small cubic neighborhood located at the first octant of a 3D Fourier space (including the DC component), are then fed into six NN during the learning stage. Each NN yields the estimates of a registration parameter. The proposed method was assessed for 3D rigid-body transformations, using DC neighborhoods of different sizes. The mean absolute registration errors are of approximately 0.030 mm in translations and 0.030 deg in rotations, for the typical motion amplitudes encountered in FMRI studies. The construction of the training set and the learning stage are fast requiring, respectively, 90 s and 1 to 12 s, depending on the number of input and hidden units of the NN. We believe that NN-based approaches to the problem of FMRI registration can be of great interest in the future. For instance, NN relying on limited K-space data (possibly in navigation echoes) can be a valid solution to the problem of prospective (in frame) FMRI registration.

Novos receptores moleculares baseados em calix[4]arenos: aplicação à química sensorial

Trabalho Final de Mestrado para obtenção do grau de Mestre em Engenharia Química

Unmixing hyperspectral intimate mixtures

Proceedings of International Conference Conference Volume 7830 Image and Signal Processing for Remote Sensing XVI Lorenzo Bruzzone Toulouse, France | September 20, 2010

Métodos "verdes" de produção de nanomateriais que promovem nanotecnologias sustentáveis

Trabalho Final de Mestrado para obtenção do grau de Mestre em Engenharia Química

A comprehensive high-throughput FTIR spectroscopy-based method for evaluating the transfection event: estimating the transfection efficiency and extracting associated metabolic responses

Reporter genes are routinely used in every laboratory for molecular and cellular biology for studying heterologous gene expression and general cellular biological mechanisms, such as transfection processes. Although well characterized and broadly implemented, reporter genes present serious limitations, either by involving time-consuming procedures or by presenting possible side effects on the expression of the heterologous gene or even in the general cellular metabolism. Fourier transform mid-infrared (FT-MIR) spectroscopy was evaluated to simultaneously analyze in a rapid (minutes) and high-throughput mode (using 96-wells microplates), the transfection efficiency, and the effect of the transfection process on the host cell biochemical composition and metabolism. Semi-adherent HEK and adherent AGS cell lines, transfected with the plasmid pVAX-GFP using Lipofectamine, were used as model systems. Good partial least squares (PLS) models were built to estimate the transfection efficiency, either considering each cell line independently (R 2 ≥ 0.92; RMSECV ≤ 2 %) or simultaneously considering both cell lines (R 2 = 0.90; RMSECV = 2 %). Additionally, the effect of the transfection process on the HEK cell biochemical and metabolic features could be evaluated directly from the FT-IR spectra. Due to the high sensitivity of the technique, it was also possible to discriminate the effect of the transfection process from the transfection reagent on KEK cells, e.g., by the analysis of spectral biomarkers and biochemical and metabolic features. The present results are far beyond what any reporter gene assay or other specific probe can offer for these purposes.

GPU implementation of the simplex identification via split augmented Lagrangian

Hyperspectral imaging can be used for object detection and for discriminating between different objects based on their spectral characteristics. One of the main problems of hyperspectral data analysis is the presence of mixed pixels, due to the low spatial resolution of such images. This means that several spectrally pure signatures (endmembers) are combined into the same mixed pixel. Linear spectral unmixing follows an unsupervised approach which aims at inferring pure spectral signatures and their material fractions at each pixel of the scene. The huge data volumes acquired by such sensors put stringent requirements on processing and unmixing methods. This paper proposes an efficient implementation of a unsupervised linear unmixing method on GPUs using CUDA. The method finds the smallest simplex by solving a sequence of nonsmooth convex subproblems using variable splitting to obtain a constraint formulation, and then applying an augmented Lagrangian technique. The parallel implementation of SISAL presented in this work exploits the GPU architecture at low level, using shared memory and coalesced accesses to memory. The results herein presented indicate that the GPU implementation can significantly accelerate the method's execution over big datasets while maintaining the methods accuracy.

Parallel GPU architecture for hyperspectral unmixing based on augmented Lagrangian method

Hyperspectral imaging has become one of the main topics in remote sensing applications, which comprise hundreds of spectral bands at different (almost contiguous) wavelength channels over the same area generating large data volumes comprising several GBs per flight. This high spectral resolution can be used for object detection and for discriminate between different objects based on their spectral characteristics. One of the main problems involved in hyperspectral analysis is the presence of mixed pixels, which arise when the spacial resolution of the sensor is not able to separate spectrally distinct materials. Spectral unmixing is one of the most important task for hyperspectral data exploitation. However, the unmixing algorithms can be computationally very expensive, and even high power consuming, which compromises the use in applications under on-board constraints. In recent years, graphics processing units (GPUs) have evolved into highly parallel and programmable systems. Specifically, several hyperspectral imaging algorithms have shown to be able to benefit from this hardware taking advantage of the extremely high floating-point processing performance, compact size, huge memory bandwidth, and relatively low cost of these units, which make them appealing for onboard data processing. In this paper, we propose a parallel implementation of an augmented Lagragian based method for unsupervised hyperspectral linear unmixing on GPUs using CUDA. The method called simplex identification via split augmented Lagrangian (SISAL) aims to identify the endmembers of a scene, i.e., is able to unmix hyperspectral data sets in which the pure pixel assumption is violated. The efficient implementation of SISAL method presented in this work exploits the GPU architecture at low level, using shared memory and coalesced accesses to memory.

Grid structure impact in sparse point representation of derivatives

In the Sparse Point Representation (SPR) method the principle is to retain the function data indicated by significant interpolatory wavelet coefficients, which are defined as interpolation errors by means of an interpolating subdivision scheme. Typically, a SPR grid is coarse in smooth regions, and refined close to irregularities. Furthermore, the computation of partial derivatives of a function from the information of its SPR content is performed in two steps. The first one is a refinement procedure to extend the SPR by the inclusion of new interpolated point values in a security zone. Then, for points in the refined grid, such derivatives are approximated by uniform finite differences, using a step size proportional to each point local scale. If required neighboring stencils are not present in the grid, the corresponding missing point values are approximated from coarser scales using the interpolating subdivision scheme. Using the cubic interpolation subdivision scheme, we demonstrate that such adaptive finite differences can be formulated in terms of a collocation scheme based on the wavelet expansion associated to the SPR. For this purpose, we prove some results concerning the local behavior of such wavelet reconstruction operators, which stand for SPR grids having appropriate structures. This statement implies that the adaptive finite difference scheme and the one using the step size of the finest level produce the same result at SPR grid points. Consequently, in addition to the refinement strategy, our analysis indicates that some care must be taken concerning the grid structure, in order to keep the truncation error under a certain accuracy limit. Illustrating results are presented for 2D Maxwell's equation numerical solutions.

Nonlinear optimization method for short-term hydro scheduling considering head-dependency

This paper is on the problem of short-term hydro scheduling, particularly concerning head-dependent reservoirs under competitive environment. We propose a new nonlinear optimization method to consider hydroelectric power generation as a function of water discharge and also of the head. Head-dependency is considered on short-term hydro scheduling in order to obtain more realistic and feasible results. The proposed method has been applied successfully to solve a case study based on one of the main Portuguese cascaded hydro systems, providing a higher profit at a negligible additional computation time in comparison with a linear optimization method that ignores head-dependency.

A method to analyse the alignment of core values in collaborative networked organisations

Since collaborative networked organisations are usually formed by independent and heterogeneous entities, it is natural that each member holds his own set of values, and that conflicts among partners might emerge because of some misalignment of values. In contrast, it is often stated in literature that the alignment between the value systems of members involved in collaborative processes is a prerequisite for successful co-working. As a result, the issue of core value alignment in collaborative networks started to attract attention. However, methods to analyse such alignment are lacking mainly because the concept of 'alignment' in this context is still ill defined and shows a multifaceted nature. As a contribution to the area, this article introduces an approach based on causal models and graph theory for the analysis of core value alignment in collaborative networks. The potential application of the approach is then discussed in the virtual organisations' breeding environment context.

Femtosecond laser ablation of dentin

The surface morphology, structure and composition of human dentin treated with a femtosecond infrared laser (pulse duration 500 fs, wavelength 1030 nm, fluences ranging from 1 to 3 J cm(-2)) was studied by scanning electron microscopy, x-ray diffraction, x-ray photoelectron spectroscopy and Fourier transform infrared spectroscopy. The average dentin ablation threshold under these conditions was 0.6 +/- 0.2 J cm(-2) and the ablation rate achieved in the range 1 to 2 mu m/pulse for an average fluence of 3 J cm(-2). The ablation surfaces present an irregular and rugged appearance, with no significant traces of melting, deformation, cracking or carbonization. The smear layer was entirely removed by the laser treatment. For fluences only slightly higher than the ablation threshold the morphology of the laser-treated surfaces was very similar to the dentin fracture surfaces and the dentinal tubules remained open. For higher fluences, the surface was more porous and the dentin structure was partially concealed by ablation debris and a few resolidified droplets. Independently on the laser processing parameters and laser processing method used no sub-superficial cracking was observed. The dentin constitution and chemical composition was not significantly modified by the laser treatment in the processing parameter range used. In particular, the organic matter is not preferentially removed from the surface and no traces of high temperature phosphates, such as the beta-tricalcium phosphate, were observed. The achieved results are compatible with an electrostatic ablation mechanism. In conclusion, the high beam quality and short pulse duration of the ultrafast laser used should allow the accurate preparation of cavities, with negligible damage of the underlying material.

A Step-up mu-Power Converter for Solar Energy Harvesting Applications, using Hill Climbing Maximum Power Point Tracking

This paper presents a step-up micro-power converter for solar energy harvesting applications. The circuit uses a SC voltage tripler architecture, controlled by an MPPT circuit based on the Hill Climbing algorithm. This circuit was designed in a 0.13 mu m CMOS technology in order to work with an a-Si PV cell. The circuit has a local power supply voltage, created using a scaled down SC voltage tripler, controlled by the same MPPT circuit, to make the circuit robust to load and illumination variations. The SC circuits use a combination of PMOS and NMOS transistors to reduce the occupied area. A charge re-use scheme is used to compensate the large parasitic capacitors associated to the MOS transistors. The simulation results show that the circuit can deliver a power of 1266 mu W to the load using 1712 mu W of power from the PV cell, corresponding to an efficiency as high as 73.91%. The simulations also show that the circuit is capable of starting up with only 19% of the maximum illumination level.

Direct multisearch for multiobjective optimization

In practical applications of optimization it is common to have several conflicting objective functions to optimize. Frequently, these functions are subject to noise or can be of black-box type, preventing the use of derivative-based techniques. We propose a novel multiobjective derivative-free methodology, calling it direct multisearch (DMS), which does not aggregate any of the objective functions. Our framework is inspired by the search/poll paradigm of direct-search methods of directional type and uses the concept of Pareto dominance to maintain a list of nondominated points (from which the new iterates or poll centers are chosen). The aim of our method is to generate as many points in the Pareto front as possible from the polling procedure itself, while keeping the whole framework general enough to accommodate other disseminating strategies, in particular, when using the (here also) optional search step. DMS generalizes to multiobjective optimization (MOO) all direct-search methods of directional type. We prove under the common assumptions used in direct search for single objective optimization that at least one limit point of the sequence of iterates generated by DMS lies in (a stationary form of) the Pareto front. However, extensive computational experience has shown that our methodology has an impressive capability of generating the whole Pareto front, even without using a search step. Two by-products of this paper are (i) the development of a collection of test problems for MOO and (ii) the extension of performance and data profiles to MOO, allowing a comparison of several solvers on a large set of test problems, in terms of their efficiency and robustness to determine Pareto fronts.