Title :
A new image correcting method based on MWLS_SVR
Author :
Dingxue, Wu ; Daiqiang, Peng ; Jinwen, Tian
Author_Institution :
Comput. Dept., Huanggang Normal Univ., Huanggang, China
Abstract :
Image correction is to establish the correspondence between two images or to determine the geometric transformation. In this paper, we present a novel method to correct image geometric distortion. The method can find the desired transformation by a set of ground control points (GCPs) based on local adaptive multi-scale wavelet least squares support vector regression(MWLS_SVR) and computer the unknown pixels location by the estimated transformations. To test the proposed method, a number of experiments are carried out. The simulated results illustrate that the proposed method can actually correct image geometric distortion and regress arbitrary distortion function such as linear and nonlinear transformation. It is demonstrated in the experiment that the performance of the method has interesting result.
Keywords :
geometry; image processing; wavelet transforms; geometric transformation; ground control point; image correcting method; image geometric distortion; multiscale wavelet least squares support vector regression; nonlinear transformation; Cities and towns; Equations; Error correction; Least squares approximation; Mathematical model; Nonlinear distortion; Pixel; Polynomials; Solid modeling; Transforms; GCPs; Geometric distortion; MWLS_SVR;
Conference_Titel :
Computer Science and Information Technology, 2009. ICCSIT 2009. 2nd IEEE International Conference on
Conference_Location :
Beijing
Print_ISBN :
978-1-4244-4519-6
Electronic_ISBN :
978-1-4244-4520-2
DOI :
10.1109/ICCSIT.2009.5234779