Font Size: a A A

The Research Of Image Denoising And Emhancement Via Contourlet Transform

Posted on:2007-06-13Degree:MasterType:Thesis
Country:ChinaCandidate:Q Y ZhaoFull Text:PDF
GTID:2178360215470189Subject:Mathematics
Abstract/Summary:PDF Full Text Request
Recent years, the development of multiscale geometric analysis beyond the wavelet transform, which has taken a great deal of success. Many scholars are studying this topic.This paper study the idea of multiscale geometric analysis, which could capture the intrinsic geometrical structure that is key in visual information, finding a kind of optimal representation of high dimension function in the sense of nonlinear approximation. Then we study the contourlet transform thoroughly. The paper begins with a detailed study on the statistic of contourlet coefficients, which reveals the relationship between the coefficients with same scale same space and different direction, we conclude the MAP estimator, developing the method of image denoising by contourlet with changed Bivariate shrinkage. Compare to the BayeShrink and BivaShrink based on wavelet, the results reveal that proposed method can get higher PSNR and the better visual. Meantime, this paper studies the problem of image enhancement with the contourlet, the results show that it is better than the method with wavelet.
Keywords/Search Tags:Wavelet transform, Multiscale geometric analysis, Contourlet transform, Image denosing, Image enhancement
PDF Full Text Request
Related items