Learning wind fields with multiple kernels

This paper presents multiple kernel learning (MKL) regression as an exploratory spatial data analysis and modelling tool. The MKL approach is introduced as an extension of support vector regression, where MKL uses dedicated kernels to divide a given task into sub-problems and to treat them separately in an effective way. It provides better interpretability to non-linear robust kernel regression at the cost of a more complex numerical optimization. In particular, we investigate the use of MKL as a tool that allows us to avoid using ad-hoc topographic indices as covariables in statistical models in complex terrains. Instead, MKL learns these relationships from the data in a non-parametric fashion. A study on data simulated from real terrain features confirms the ability of MKL to enhance the interpretability of data-driven models and to aid feature selection without degrading predictive performances. Here we examine the stability of the MKL algorithm with respect to the number of training data samples and to the presence of noise. The results of a real case study are also presented, where MKL is able to exploit a large set of terrain features computed at multiple spatial scales, when predicting mean wind speed in an Alpine region.

Multiple Kernel Learning of Environmental Data. Case study: Analysis and Mapping of Wind Fields

The paper presents the Multiple Kernel Learning (MKL) approach as a modelling and data exploratory tool and applies it to the problem of wind speed mapping. Support Vector Regression (SVR) is used to predict spatial variations of the mean wind speed from terrain features (slopes, terrain curvature, directional derivatives) generated at different spatial scales. Multiple Kernel Learning is applied to learn kernels for individual features and thematic feature subsets, both in the context of feature selection and optimal parameters determination. An empirical study on real-life data confirms the usefulness of MKL as a tool that enhances the interpretability of data-driven models.

Fast Geodesic Active Fields for Image Registration Based on Splitting and Augmented Lagrangian Approaches.

In this paper, we present an efficient numerical scheme for the recently introduced geodesic active fields (GAF) framework for geometric image registration. This framework considers the registration task as a weighted minimal surface problem. Hence, the data-term and the regularization-term are combined through multiplication in a single, parametrization invariant and geometric cost functional. The multiplicative coupling provides an intrinsic, spatially varying and data-dependent tuning of the regularization strength, and the parametrization invariance allows working with images of nonflat geometry, generally defined on any smoothly parametrizable manifold. The resulting energy-minimizing flow, however, has poor numerical properties. Here, we provide an efficient numerical scheme that uses a splitting approach; data and regularity terms are optimized over two distinct deformation fields that are constrained to be equal via an augmented Lagrangian approach. Our approach is more flexible than standard Gaussian regularization, since one can interpolate freely between isotropic Gaussian and anisotropic TV-like smoothing. In this paper, we compare the geodesic active fields method with the popular Demons method and three more recent state-of-the-art algorithms: NL-optical flow, MRF image registration, and landmark-enhanced large displacement optical flow. Thus, we can show the advantages of the proposed FastGAF method. It compares favorably against Demons, both in terms of registration speed and quality. Over the range of example applications, it also consistently produces results not far from more dedicated state-of-the-art methods, illustrating the flexibility of the proposed framework.

Kernel-based Mapping of Meteorological Fields in Complex Orography

Due to the advances in sensor networks and remote sensing technologies, the acquisition and storage rates of meteorological and climatological data increases every day and ask for novel and efficient processing algorithms. A fundamental problem of data analysis and modeling is the spatial prediction of meteorological variables in complex orography, which serves among others to extended climatological analyses, for the assimilation of data into numerical weather prediction models, for preparing inputs to hydrological models and for real time monitoring and short-term forecasting of weather.In this thesis, a new framework for spatial estimation is proposed by taking advantage of a class of algorithms emerging from the statistical learning theory. Nonparametric kernel-based methods for nonlinear data classification, regression and target detection, known as support vector machines (SVM), are adapted for mapping of meteorological variables in complex orography.With the advent of high resolution digital elevation models, the field of spatial prediction met new horizons. In fact, by exploiting image processing tools along with physical heuristics, an incredible number of terrain features which account for the topographic conditions at multiple spatial scales can be extracted. Such features are highly relevant for the mapping of meteorological variables because they control a considerable part of the spatial variability of meteorological fields in the complex Alpine orography. For instance, patterns of orographic rainfall, wind speed and cold air pools are known to be correlated with particular terrain forms, e.g. convex/concave surfaces and upwind sides of mountain slopes.Kernel-based methods are employed to learn the nonlinear statistical dependence which links the multidimensional space of geographical and topographic explanatory variables to the variable of interest, that is the wind speed as measured at the weather stations or the occurrence of orographic rainfall patterns as extracted from sequences of radar images. Compared to low dimensional models integrating only the geographical coordinates, the proposed framework opens a way to regionalize meteorological variables which are multidimensional in nature and rarely show spatial auto-correlation in the original space making the use of classical geostatistics tangled.The challenges which are explored during the thesis are manifolds. First, the complexity of models is optimized to impose appropriate smoothness properties and reduce the impact of noisy measurements. Secondly, a multiple kernel extension of SVM is considered to select the multiscale features which explain most of the spatial variability of wind speed. Then, SVM target detection methods are implemented to describe the orographic conditions which cause persistent and stationary rainfall patterns. Finally, the optimal splitting of the data is studied to estimate realistic performances and confidence intervals characterizing the uncertainty of predictions.The resulting maps of average wind speeds find applications within renewable resources assessment and opens a route to decrease the temporal scale of analysis to meet hydrological requirements. Furthermore, the maps depicting the susceptibility to orographic rainfall enhancement can be used to improve current radar-based quantitative precipitation estimation and forecasting systems and to generate stochastic ensembles of precipitation fields conditioned upon the orography.

Exposure to radio-frequency electromagnetic fields from broadcast transmitters and risk of childhood cancer: a census-based cohort study.

We investigated the association between exposure to radio-frequency electromagnetic fields (RF-EMFs) from broadcast transmitters and childhood cancer. First, we conducted a time-to-event analysis including children under age 16 years living in Switzerland on December 5, 2000. Follow-up lasted until December 31, 2008. Second, all children living in Switzerland for some time between 1985 and 2008 were included in an incidence density cohort. RF-EMF exposure from broadcast transmitters was modeled. Based on 997 cancer cases, adjusted hazard ratios in the time-to-event analysis for the highest exposure category (>0.2 V/m) as compared with the reference category (<0.05 V/m) were 1.03 (95% confidence interval (CI): 0.74, 1.43) for all cancers, 0.55 (95% CI: 0.26, 1.19) for childhood leukemia, and 1.68 (95% CI: 0.98, 2.91) for childhood central nervous system (CNS) tumors. Results of the incidence density analysis, based on 4,246 cancer cases, were similar for all types of cancer and leukemia but did not indicate a CNS tumor risk (incidence rate ratio = 1.03, 95% CI: 0.73, 1.46). This large census-based cohort study did not suggest an association between predicted RF-EMF exposure from broadcasting and childhood leukemia. Results for CNS tumors were less consistent, but the most comprehensive analysis did not suggest an association.

Active deformation fields: dense deformation field estimation for atlas-based segmentation using the active contour framework.

This paper presents a new and original variational framework for atlas-based segmentation. The proposed framework integrates both the active contour framework, and the dense deformation fields of optical flow framework. This framework is quite general and encompasses many of the state-of-the-art atlas-based segmentation methods. It also allows to perform the registration of atlas and target images based on only selected structures of interest. The versatility and potentiality of the proposed framework are demonstrated by presenting three diverse applications: In the first application, we show how the proposed framework can be used to simulate the growth of inconsistent structures like a tumor in an atlas. In the second application, we estimate the position of nonvisible brain structures based on the surrounding structures and validate the results by comparing with other methods. In the final application, we present the segmentation of lymph nodes in the Head and Neck CT images, and demonstrate how multiple registration forces can be used in this framework in an hierarchical manner.

Minimization of Nyquist ghosting for echo-planar imaging at ultra-high fields based on a "negative readout gradient" strategy.

PURPOSE: To improve the traditional Nyquist ghost correction approach in echo planar imaging (EPI) at high fields, via schemes based on the reversal of the EPI readout gradient polarity for every other volume throughout a functional magnetic resonance imaging (fMRI) acquisition train. MATERIALS AND METHODS: An EPI sequence in which the readout gradient was inverted every other volume was implemented on two ultrahigh-field systems. Phantom images and fMRI data were acquired to evaluate ghost intensities and the presence of false-positive blood oxygenation level-dependent (BOLD) signal with and without ghost correction. Three different algorithms for ghost correction of alternating readout EPI were compared. RESULTS: Irrespective of the chosen processing approach, ghosting was significantly reduced (up to 70% lower intensity) in both rat brain images acquired on a 9.4T animal scanner and human brain images acquired at 7T, resulting in a reduction of sources of false-positive activation in fMRI data. CONCLUSION: It is concluded that at high B(0) fields, substantial gains in Nyquist ghost correction of echo planar time series are possible by alternating the readout gradient every other volume.

On the convergence of EM-like algorithms for image segmentation using Markov random fields.

Inference of Markov random field images segmentation models is usually performed using iterative methods which adapt the well-known expectation-maximization (EM) algorithm for independent mixture models. However, some of these adaptations are ad hoc and may turn out numerically unstable. In this paper, we review three EM-like variants for Markov random field segmentation and compare their convergence properties both at the theoretical and practical levels. We specifically advocate a numerical scheme involving asynchronous voxel updating, for which general convergence results can be established. Our experiments on brain tissue classification in magnetic resonance images provide evidence that this algorithm may achieve significantly faster convergence than its competitors while yielding at least as good segmentation results.

In vivo Structural Imaging of the Cerebellum, the Contribution of Ultra-High Fields.

This review covers some of the contributions to date from cerebellar imaging studies performed at ultra-high magnetic fields. A short overview of the general advantages and drawbacks of the use of such high field systems for imaging is given. One of the biggest advantages of imaging at high magnetic fields is the improved spatial resolution, achievable thanks to the increased available signal-to-noise ratio. This high spatial resolution better matches the dimensions of the cerebellar substructures, allowing a better definition of such structures in the images. The implications of the use of high field systems is discussed for several imaging sequences and image contrast mechanisms. This review covers studies which were performed in vivo in both rodents and humans, with a special focus on studies that were directed towards the observation of the different cerebellar layers.

The Effect of Two Speed Endurance Training Regimes on Performance of Soccer Players.

In order to better understand the specificity of training adaptations, we compared the effects of two different anaerobic training regimes on various types of soccer-related exercise performances. During the last 3 weeks of the competitive season, thirteen young male professional soccer players (age 18.5±1 yr, height 179.5±6.5 cm, body mass 74.3±6.5 kg) reduced the training volume by ~20% and replaced their habitual fitness conditioning work with either speed endurance production (SEP; n = 6) or speed endurance maintenance (SEM; n = 7) training, three times per wk. SEP training consisted of 6-8 reps of 20-s all-out running bouts followed by 2 min of passive recovery, whereas SEM training was characterized by 6-8 x 20-s all-out efforts interspersed with 40 s of passive recovery. SEP training reduced (p<0.01) the total time in a repeated sprint ability test (RSAt) by 2.5%. SEM training improved the 200-m sprint performance (from 26.59±0.70 to 26.02±0.62 s, p<0.01) and had a likely beneficial impact on the percentage decrement score of the RSA test (from 4.07±1.28 to 3.55±1.01%) but induced a very likely impairment in RSAt (from 83.81±2.37 to 84.65±2.27 s). The distance covered in the Yo-Yo Intermittent Recovery test level 2 was 10.1% (p<0.001) and 3.8% (p<0.05) higher after SEP and SEM training, respectively, with possibly greater improvements following SEP compared to SEM. No differences were observed in the 20- and 40-m sprint performances. In conclusion, these two training strategies target different determinants of soccer-related physical performance. SEP improved repeated sprint and high-intensity intermittent exercise performance, whereas SEM increased muscles' ability to maximize fatigue tolerance and maintain speed development during both repeated all-out and continuous short-duration maximal exercises. These results provide new insight into the precise nature of a stimulus necessary to improve specific types of athletic performance in trained young soccer players.

Moral Disengagement in the Legitimation and Realization of Aggressive Behaviors in Soccer and Ice Hockey

The aim of the present study was to test the effect of moral disengagement on the tolerance and realization of aggressive acts in male soccer and ice hockey players in Switzerland. One hundred and four soccer and 98 ice hockey players evaluated the legitimacy of four videotaped aggressive behaviors and completed a questionnaire that included a moral disengagement scale and self-reported aggression. The level of moral disengagement, which mediates the effects of perceived coach and ego attitudes toward transgressions, largely explains the tolerance of hostile aggression within teams, as well as the level of high aggressive acts reported by the participants.

Maintenance Therapy With Tumor-Treating Fields Plus Temozolomide vs Temozolomide Alone for Glioblastoma: A Randomized Clinical Trial.

IMPORTANCE: Glioblastoma is the most devastating primary malignancy of the central nervous system in adults. Most patients die within 1 to 2 years of diagnosis. Tumor-treating fields (TTFields) are a locoregionally delivered antimitotic treatment that interferes with cell division and organelle assembly. OBJECTIVE: To evaluate the efficacy and safety of TTFields used in combination with temozolomide maintenance treatment after chemoradiation therapy for patients with glioblastoma. DESIGN, SETTING, AND PARTICIPANTS: After completion of chemoradiotherapy, patients with glioblastoma were randomized (2:1) to receive maintenance treatment with either TTFields plus temozolomide (n = 466) or temozolomide alone (n = 229) (median time from diagnosis to randomization, 3.8 months in both groups). The study enrolled 695 of the planned 700 patients between July 2009 and November 2014 at 83 centers in the United States, Canada, Europe, Israel, and South Korea. The trial was terminated based on the results of this planned interim analysis. INTERVENTIONS: Treatment with TTFields was delivered continuously (>18 hours/day) via 4 transducer arrays placed on the shaved scalp and connected to a portable medical device. Temozolomide (150-200 mg/m2/d) was given for 5 days of each 28-day cycle. MAIN OUTCOMES AND MEASURES: The primary end point was progression-free survival in the intent-to-treat population (significance threshold of .01) with overall survival in the per-protocol population (n = 280) as a powered secondary end point (significance threshold of .006). This prespecified interim analysis was to be conducted on the first 315 patients after at least 18 months of follow-up. RESULTS: The interim analysis included 210 patients randomized to TTFields plus temozolomide and 105 randomized to temozolomide alone, and was conducted at a median follow-up of 38 months (range, 18-60 months). Median progression-free survival in the intent-to-treat population was 7.1 months (95% CI, 5.9-8.2 months) in the TTFields plus temozolomide group and 4.0 months (95% CI, 3.3-5.2 months) in the temozolomide alone group (hazard ratio [HR], 0.62 [98.7% CI, 0.43-0.89]; P = .001). Median overall survival in the per-protocol population was 20.5 months (95% CI, 16.7-25.0 months) in the TTFields plus temozolomide group (n = 196) and 15.6 months (95% CI, 13.3-19.1 months) in the temozolomide alone group (n = 84) (HR, 0.64 [99.4% CI, 0.42-0.98]; P = .004). CONCLUSIONS AND RELEVANCE: In this interim analysis of 315 patients with glioblastoma who had completed standard chemoradiation therapy, adding TTFields to maintenance temozolomide chemotherapy significantly prolonged progression-free and overall survival. TRIAL REGISTRATION: clinicaltrials.gov Identifier: NCT00916409.