855 resultados para Asynchronous iterative algorithms
Resumo:
The aim of this study is to perform a thorough comparison of quantitative susceptibility mapping (QSM) techniques and their dependence on the assumptions made. The compared methodologies were: two iterative single orientation methodologies minimizing the l2, l1TV norm of the prior knowledge of the edges of the object, one over-determined multiple orientation method (COSMOS) and anewly proposed modulated closed-form solution (MCF). The performance of these methods was compared using a numerical phantom and in-vivo high resolution (0.65mm isotropic) brain data acquired at 7T using a new coil combination method. For all QSM methods, the relevant regularization and prior-knowledge parameters were systematically changed in order to evaluate the optimal reconstruction in the presence and absence of a ground truth. Additionally, the QSM contrast was compared to conventional gradient recalled echo (GRE) magnitude and R2* maps obtained from the same dataset. The QSM reconstruction results of the single orientation methods show comparable performance. The MCF method has the highest correlation (corrMCF=0.95, r(2)MCF =0.97) with the state of the art method (COSMOS) with additional advantage of extreme fast computation time. The l-curve method gave the visually most satisfactory balance between reduction of streaking artifacts and over-regularization with the latter being overemphasized when the using the COSMOS susceptibility maps as ground-truth. R2* and susceptibility maps, when calculated from the same datasets, although based on distinct features of the data, have a comparable ability to distinguish deep gray matter structures.
Resumo:
Defining an efficient training set is one of the most delicate phases for the success of remote sensing image classification routines. The complexity of the problem, the limited temporal and financial resources, as well as the high intraclass variance can make an algorithm fail if it is trained with a suboptimal dataset. Active learning aims at building efficient training sets by iteratively improving the model performance through sampling. A user-defined heuristic ranks the unlabeled pixels according to a function of the uncertainty of their class membership and then the user is asked to provide labels for the most uncertain pixels. This paper reviews and tests the main families of active learning algorithms: committee, large margin, and posterior probability-based. For each of them, the most recent advances in the remote sensing community are discussed and some heuristics are detailed and tested. Several challenging remote sensing scenarios are considered, including very high spatial resolution and hyperspectral image classification. Finally, guidelines for choosing the good architecture are provided for new and/or unexperienced user.
Resumo:
"Vegeu el resum a l'inici del document del fitxer adjunt."
Resumo:
In a seminal paper [10], Weitz gave a deterministic fully polynomial approximation scheme for counting exponentially weighted independent sets (which is the same as approximating the partition function of the hard-core model from statistical physics) in graphs of degree at most d, up to the critical activity for the uniqueness of the Gibbs measure on the innite d-regular tree. ore recently Sly [8] (see also [1]) showed that this is optimal in the sense that if here is an FPRAS for the hard-core partition function on graphs of maximum egree d for activities larger than the critical activity on the innite d-regular ree then NP = RP. In this paper we extend Weitz's approach to derive a deterministic fully polynomial approximation scheme for the partition function of general two-state anti-ferromagnetic spin systems on graphs of maximum degree d, up to the corresponding critical point on the d-regular tree. The main ingredient of our result is a proof that for two-state anti-ferromagnetic spin systems on the d-regular tree, weak spatial mixing implies strong spatial mixing. his in turn uses a message-decay argument which extends a similar approach proposed recently for the hard-core model by Restrepo et al [7] to the case of general two-state anti-ferromagnetic spin systems.
Resumo:
Report for the scientific sojourn at the University of California at Berkeley, USA, from september 2007 until july 2008. Communities of Learning Practice is an innovative paradigm focused on providing appropriate technological support to both formal and especially informal learning groups who are chiefly formed by non-technical people and who lack of the necessary resources to acquire such systems. Typically, students who are often separated by geography and/or time have the need to meet each other after classes in small study groups to carry out specific learning activities assigned during the formal learning process. However, the lack of suitable and available groupware applications makes it difficult for these groups of learners to collaborate and achieve their specific learning goals. In addition, the lack of democratic decision-making mechanisms is a main handicap to substitute the central authority of knowledge presented in formal learning.
Resumo:
The paper presents an approach for mapping of precipitation data. The main goal is to perform spatial predictions and simulations of precipitation fields using geostatistical methods (ordinary kriging, kriging with external drift) as well as machine learning algorithms (neural networks). More practically, the objective is to reproduce simultaneously both the spatial patterns and the extreme values. This objective is best reached by models integrating geostatistics and machine learning algorithms. To demonstrate how such models work, two case studies have been considered: first, a 2-day accumulation of heavy precipitation and second, a 6-day accumulation of extreme orographic precipitation. The first example is used to compare the performance of two optimization algorithms (conjugate gradients and Levenberg-Marquardt) of a neural network for the reproduction of extreme values. Hybrid models, which combine geostatistical and machine learning algorithms, are also treated in this context. The second dataset is used to analyze the contribution of radar Doppler imagery when used as external drift or as input in the models (kriging with external drift and neural networks). Model assessment is carried out by comparing independent validation errors as well as analyzing data patterns.
Resumo:
This paper characterizes and evaluates the potential of three commercial CT iterative reconstruction methods (ASIR?, VEO? and iDose(4 ()?())) for dose reduction and image quality improvement. We measured CT number accuracy, standard deviation (SD), noise power spectrum (NPS) and modulation transfer function (MTF) metrics on Catphan phantom images while five human observers performed four-alternative forced-choice (4AFC) experiments to assess the detectability of low- and high-contrast objects embedded in two pediatric phantoms. Results show that 40% and 100% ASIR as well as iDose(4) levels 3 and 6 do not affect CT number and strongly decrease image noise with relative SD constant in a large range of dose. However, while ASIR produces a shift of the NPS curve apex, less change is observed with iDose(4) with respect to FBP methods. With second-generation iterative reconstruction VEO, physical metrics are even further improved: SD decreased to 70.4% at 0.5 mGy and spatial resolution improved to 37% (MTF(50%)). 4AFC experiments show that few improvements in detection task performance are obtained with ASIR and iDose(4), whereas VEO makes excellent detections possible even at an ultra-low-dose (0.3 mGy), leading to a potential dose reduction of a factor 3 to 7 (67%-86%). In spite of its longer reconstruction time and the fact that clinical studies are still required to complete these results, VEO clearly confirms the tremendous potential of iterative reconstructions for dose reduction in CT and appears to be an important tool for patient follow-up, especially for pediatric patients where cumulative lifetime dose still remains high.
Resumo:
RATIONALE AND OBJECTIVES: Dose reduction may compromise patients because of a decrease of image quality. Therefore, the amount of dose savings in new dose-reduction techniques needs to be thoroughly assessed. To avoid repeated studies in one patient, chest computed tomography (CT) scans with different dose levels were performed in corpses comparing model-based iterative reconstruction (MBIR) as a tool to enhance image quality with current standard full-dose imaging. MATERIALS AND METHODS: Twenty-five human cadavers were scanned (CT HD750) after contrast medium injection at different, decreasing dose levels D0-D5 and respectively reconstructed with MBIR. The data at full-dose level, D0, have been additionally reconstructed with standard adaptive statistical iterative reconstruction (ASIR), which represented the full-dose baseline reference (FDBR). Two radiologists independently compared image quality (IQ) in 3-mm multiplanar reformations for soft-tissue evaluation of D0-D5 to FDBR (-2, diagnostically inferior; -1, inferior; 0, equal; +1, superior; and +2, diagnostically superior). For statistical analysis, the intraclass correlation coefficient (ICC) and the Wilcoxon test were used. RESULTS: Mean CT dose index values (mGy) were as follows: D0/FDBR = 10.1 ± 1.7, D1 = 6.2 ± 2.8, D2 = 5.7 ± 2.7, D3 = 3.5 ± 1.9, D4 = 1.8 ± 1.0, and D5 = 0.9 ± 0.5. Mean IQ ratings were as follows: D0 = +1.8 ± 0.2, D1 = +1.5 ± 0.3, D2 = +1.1 ± 0.3, D3 = +0.7 ± 0.5, D4 = +0.1 ± 0.5, and D5 = -1.2 ± 0.5. All values demonstrated a significant difference to baseline (P < .05), except mean IQ for D4 (P = .61). ICC was 0.91. CONCLUSIONS: Compared to ASIR, MBIR allowed for a significant dose reduction of 82% without impairment of IQ. This resulted in a calculated mean effective dose below 1 mSv.
Resumo:
This paper presents general problems and approaches for the spatial data analysis using machine learning algorithms. Machine learning is a very powerful approach to adaptive data analysis, modelling and visualisation. The key feature of the machine learning algorithms is that they learn from empirical data and can be used in cases when the modelled environmental phenomena are hidden, nonlinear, noisy and highly variable in space and in time. Most of the machines learning algorithms are universal and adaptive modelling tools developed to solve basic problems of learning from data: classification/pattern recognition, regression/mapping and probability density modelling. In the present report some of the widely used machine learning algorithms, namely artificial neural networks (ANN) of different architectures and Support Vector Machines (SVM), are adapted to the problems of the analysis and modelling of geo-spatial data. Machine learning algorithms have an important advantage over traditional models of spatial statistics when problems are considered in a high dimensional geo-feature spaces, when the dimension of space exceeds 5. Such features are usually generated, for example, from digital elevation models, remote sensing images, etc. An important extension of models concerns considering of real space constrains like geomorphology, networks, and other natural structures. Recent developments in semi-supervised learning can improve modelling of environmental phenomena taking into account on geo-manifolds. An important part of the study deals with the analysis of relevant variables and models' inputs. This problem is approached by using different feature selection/feature extraction nonlinear tools. To demonstrate the application of machine learning algorithms several interesting case studies are considered: digital soil mapping using SVM, automatic mapping of soil and water system pollution using ANN; natural hazards risk analysis (avalanches, landslides), assessments of renewable resources (wind fields) with SVM and ANN models, etc. The dimensionality of spaces considered varies from 2 to more than 30. Figures 1, 2, 3 demonstrate some results of the studies and their outputs. Finally, the results of environmental mapping are discussed and compared with traditional models of geostatistics.