Search results for "Regularization"
showing 10 items of 189 documents
Regularization Method in Infrared Image Processing
2003
Abstract Infrared images often present distortions induced by the measurement system. Thus, image processing is a vital part of infrared measurements. A distortion model based on a convolution product is presented. Image restoration is an ill-posed problem and its solution can be obtained using regularization methods. In this paper, 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 matrix that admits a diagonal form in the two-dimensional Fourier space. The restoration procedure is used to restore a knife-edge infrared source image.
New Regularization Method for EXAFS Analysis
2007
As an alternative to the analysis of EXAFS spectra by conventional shell fitting, the Tikhonov regularization method has been proposed. An improved algorithm that utilizes a priori information about the sample has been developed and applied to the analysis of U L3‐edge spectra of soddyite, (UO2)2SiO4⋅2H2O, and of U(VI) sorbed onto kaolinite. The partial radial distribution functions g1(UU), g2(USi), and g3(UO) of soddyite agree with crystallographic values and previous EXAFS results.
Deconvolution of Multiple Spectral Lines Shapes by Means of Tikhonov’s Regularization Method
2013
We present deconvolution of multiple narrow Zeeman split Hg lines, emitted from Hg/Xe micro-size capillary and measured by the Fourier Transform spectrometer. The ill-posed inverse problem was solved using the Tikhonov& rsquo;s regularization method.
MR3714763 Reviewed Bargetz, C.(A-INSB); Nigsch, E. A.(A-WIEN-WPI); Ortner, N.(A-INSB) Convolvability and regularization of distributions. (English su…
2018
Referring to the theory of vector-valued distributions due to L. Schwartz, the authors, starting from a formulation due to Hirata and Shiraishi, carry out a study about generalizations of the convolvability and regularization of distributions, without test functions but by means of kernels. Further topological features, such as boundedness and relative compactness of subsets of distributions, are exhibited in light of previous results.
ELM Regularized Method for Classification Problems
2016
Extreme Learning Machine (ELM) is a recently proposed algorithm, efficient and fast for learning the parameters of single layer neural structures. One of the main problems of this algorithm is to choose the optimal architecture for a given problem solution. To solve this limitation several solutions have been proposed in the literature, including the regularization of the structure. However, to the best of our knowledge, there are no works where such adjustment is applied to classification problems in the presence of a non-linearity in the output; all published works tackle modelling or regression problems. Our proposal has been applied to a series of standard databases for the evaluation o…
Fast nonstationary preconditioned iterative methods for ill-posed problems, with application to image deblurring
2013
We introduce a new iterative scheme for solving linear ill-posed problems, similar to nonstationary iterated Tikhonov regularization, but with an approximation of the underlying operator to be used for the Tikhonov equations. For image deblurring problems, such an approximation can be a discrete deconvolution that operates entirely in the Fourier domain. We provide a theoretical analysis of the new scheme, using regularization parameters that are chosen by a certain adaptive strategy. The numerical performance of this method turns out to be superior to state-of-the-art iterative methods, including the conjugate gradient iteration for the normal equation, with and without additional precondi…
On the discrete linear ill‐posed problems
1999
An inverse problem of photo‐acoustic spectroscopy of semiconductors is investigated. The main problem is formulated as the integral equation of the first kind. Two different regularization methods are applied, the algorithms for defining regularization parameters are given. Diskrečiųjų blogai sąlygotų uždavinių klausimu Santrauka Darbe nagrinejamas foto‐akustines spektroskopijos puslaidininkiuose uždavinys, kuriame i vertinami nešeju difuzijos ir rekombinacijos procesai. Reikia atstatyti šaltinio funkcija f(x), jei žinoma antrosios eiles difuzijos lygtis ir atitinkamos kraštines salygos. Naudojantis matavimu, atliktu ivairiuose dažniuose, rezultatais sprendžiamas atvirkštinis uždavinys, kel…
Multicomponent line profile restoring by means of ill-posed inverse task solution
2017
The investigation of the criteria of usage of the Tikhonov regularization method for multicomponent overlapping line profiles restoring was done by means of model task solution. The influence of the width and kind of the instrumental function, number of the components of the profile and distance between components are discussed.
Differential inclusions involving normal cones of nonregular sets in Hilbert spaces
2017
This thesis is dedicated to the study of differential inclusions involving normal cones of nonregular sets in Hilbert spaces. In particular, we are interested in the sweeping process and its variants. The sweeping process is a constrained differential inclusion involving normal cones which appears naturally in several applications such as elastoplasticity, electrical circuits, hysteresis, crowd motion, etc.This work is divided conceptually in three parts: Study of positively alpha-far sets, existence results for differential inclusions involving normal cones and characterizations of Lyapunov pairs for the sweeping process. In the first part (Chapter 2), we investigate the class of positivel…
Evaluation of the areal material distribution of paper from its optical transmission image
2011
International audience; The goal of this study was to evaluate the areal mass distribution (defined as the X-ray transmission image) of paper from its optical transmission image. A Bayesian inversion framework was used in the related deconvolution process so as to combine indirect optical information with a priori knowledge about the type of paper imaged. The a priori knowledge was expressed in the form of an empirical Besov space prior distribution constructed in a computationally effective way using the wavelet transform. The estimation process took the form of a large-scale optimization problem, which was in turn solved using the gradient descent method of Barzilai and Borwein. It was de…