Image Enhancement By Nonlinear Wavelet Processing英文电子书.pdf

Image enhancement by nonlinear wavelet pro cessing J E Odegard Department of Electrical and Computer Engineering MS Rice University Houston TX email o degardriceedu w httpjazzriceedu Abstract In this pap er we describ e how the theory of wavelet thresholding intro duced by Donoho and Johnstone can successfully b e applied to two distinct problems in image pro cessing where traditional linear ltering techniques are insucient The rst application is related to sp eckle reduction in coherent imaging systems We show that the prop osed metho d works well for reducing sp eckle in SAR images while maintaining bright reections for subsequent pro cessing and detection Secondly we apply the wavelet based metho d for reducing blo cking artifacts asso ciated with most DCT based image co ders eg most notably the Joint Photographic Exp erts Group JPEG standard at high compression ratios In particular we demonstrate an algorithm for p ostpro cessing deco ded images without the need for a novel co derdeco der By applying this algorithm we are able to obtain p erceptually sup erior images at high compression ratios using the JPEG co ding standard For b oth applications we h ave develop ed metho ds for estimating the required threshold parameter and we h ave applied these to large numb er of images to study the eect of the wavelet thresholding Our main goal with this pap er is to illustrate how the recent theory of wavelet denoising can b e applied to a wide range of practical problems which do es not necessarily satisfy all the assumptions of the develop ed theory Intro duction Recent developments in wavelet theory