Search results for "Regularization"
showing 10 items of 189 documents
Large scale semi-supervised image segmentation with active queries
2011
A semiautomatic procedure to generate classification maps of remote sensing images is proposed. Starting from a hierarchical unsupervised classification, the algorithm exploits the few available labeled pixels to assign each cluster to the most probable class. For a given amount of labeled pixels, the algorithm returns a classified segmentation map, along with confidence levels of class membership for each pixel. Active learning methods are used to select the most informative samples to increase confidence in the class membership. Experiments on a AVIRIS hyperspectral image confirm the effectiveness of the method, especially when used with active learning query functions and spatial regular…
Infrared renormalization of two-loop integrals and the chiral expansion of the nucleon mass
2007
We describe details of the renormalization of two-loop integrals relevant to the calculation of the nucleon mass in the framework of manifestly Lorentz-invariant chiral perturbation theory using infrared renormalization. It is shown that the renormalization can be performed while preserving all relevant symmetries, in particular chiral symmetry, and that renormalized diagrams respect the standard power counting rules. As an application we calculate the chiral expansion of the nucleon mass to order O(q^6).
Sequential Learning with LS-SVM for Large-Scale Data Sets
2006
We present a subspace-based variant of LS-SVMs (i.e. regularization networks) that sequentially processes the data and is hence especially suited for online learning tasks. The algorithm works by selecting from the data set a small subset of basis functions that is subsequently used to approximate the full kernel on arbitrary points. This subset is identified online from the data stream. We improve upon existing approaches (esp. the kernel recursive least squares algorithm) by proposing a new, supervised criterion for the selection of the relevant basis functions that takes into account the approximation error incurred from approximating the kernel as well as the reduction of the cost in th…
Quasi-Newton approach to nonnegative image restorations
2000
Abstract Image restoration, or deblurring, is the process of attempting to correct for degradation in a recorded image. Typically the blurring system is assumed to be linear and spatially invariant, and fast Fourier transform (FFT) based schemes result in efficient computational image restoration methods. However, real images have properties that cannot always be handled by linear methods. In particular, an image consists of positive light intensities, and thus a nonnegativity constraint should be enforced. This constraint and other ways of incorporating a priori information have been suggested in various applications, and can lead to substantial improvements in the reconstructions. Neverth…
Focal plane array infrared camera transfer function calculation and image restoration
2004
Infrared images often present distortions induced by the measurement system. Image processing is thus an essential part of infrared measurements. A distortion model based on a convolution product is presented. The analytical form of the convolution kernel has been obtained from an image formation theory, along with an analysis of the sampling of the focal plane array camera detector's matrix. Image restoration is an ill-posed problem, and its solution can be obtained using regularization methods. In this work, image restoration is performed using a variation of Tikhonov regularization that makes use of the particular form of the convolution kernel matrix, which is built as a block-circulant…
A smallest irregular oriented graph containing a given diregular one
2004
AbstractA digraph is called irregular if its vertices have mutually distinct ordered pairs of semi-degrees. Let D be any diregular oriented graph (without loops or 2-dicycles). A smallest irregular oriented graph F, F=F(D), is constructed such that F includes D as an induced subdigraph, the smallest digraph being one with smallest possible order and with smallest possible size. If the digraph D is arcless then V(D) is an independent set of F(D) comprising almost all vertices of F(D) as |V(D)|→∞. The number of irregular oriented graphs is proved to be superexponential in their order. We could not show that almost all oriented graphs are/are not irregular.
Weak regularity and consecutive topologizations and regularizations of pretopologies
2009
Abstract L. Foged proved that a weakly regular topology on a countable set is regular. In terms of convergence theory, this means that the topological reflection Tξ of a regular pretopology ξ on a countable set is regular. It is proved that this still holds if ξ is a regular σ -compact pretopology. On the other hand, it is proved that for each n ω there is a (regular) pretopology ρ (on a set of cardinality c ) such that ( RT ) k ρ > ( RT ) n ρ for each k n and ( RT ) n ρ is a Hausdorff compact topology, where R is the reflector to regular pretopologies. It is also shown that there exists a regular pretopology of Hausdorff RT -order ⩾ ω 0 . Moreover, all these pretopologies have the property…
Polynomial Smoothing Splines
2014
Interpolating splines is a perfect tool for approximation of a continuous-time signal \(f(t)\) in the case when samples \(x[k]=f(k),\;k\in \mathbb {Z}\) are available. However, frequently, the samples are corrupted by random noise. In such case, the so-called smoothing splines provide better approximation. In this chapter we describe periodic smoothing splines in one and two dimensions. The SHA technique provides explicit expression of such splines and enables us to derive optimal values of the regularization parameters.
Regularized extreme learning machine for regression problems
2011
Extreme learning machine (ELM) is a new learning algorithm for single-hidden layer feedforward networks (SLFNs) proposed by Huang et al. [1]. Its main advantage is the lower computational cost, which is especially relevant when dealing with many patterns defined in a high-dimensional space. This paper proposes an algorithm for pruning ELM networks by using regularized regression methods, thus obtaining a suitable number of the hidden nodes in the network architecture. Beginning from an initial large number of hidden nodes, irrelevant nodes are then pruned using ridge regression, elastic net and lasso methods; hence, the architectural design of ELM network can be automated. Empirical studies…
Prediction of type 2 diabetes mellitus based on nutrition data
2021
Abstract Numerous predictive models for the risk of type 2 diabetes mellitus (T2DM) exist, but a minority of them has implemented nutrition data so far, even though the significant effect of nutrition on the pathogenesis, prevention and management of T2DM has been established. Thus, in the present study, we aimed to build a predictive model for the risk of T2DM that incorporates nutrition data and calculates its predictive performance. We analysed cross-sectional data from 1591 individuals from the population-based Cooperative Health Research in the Region of Augsburg (KORA) FF4 study (2013–14) and used a bootstrap enhanced elastic net penalised multivariate regression method in order to bu…