An Efficient Way of Solving Inverse Problem Using Nonlinear Wiener Filter and its Application to Pattern Recognition

The Wiener Filter is a standard means of optimizing the linear Inverse Problem, however, the revitalization of Nonlinear Inverse Problem and its empirical error reduction has remained problematic. This paper reports a novel technique of removing noise, using an approximated Wiener Filter signal in Reproducing Kernel Hilbert Space domain. Kernel Method is one of the state of the art methods that implicitly pursue nonlinear mapping of sample data into a high dimensional vector space. In order to show the incentive of the proposed method, experiments are manipulated in denoising of images and estimating the errors. Moreover, the proposed method has more precise algorithm, higher accuracy and reduced computational complexity.