DSpace
 

EMU I-REP >
02 Faculty of Engineering >
Department of Electrical and Electronic Engineering >
Theses (Master's and Ph.D) – Electrical and Electronic Engineering >

Please use this identifier to cite or link to this item: http://hdl.handle.net/11129/57

Title: A Design of Q-shift Filter for Dual-Tree Complex Wavelet Transforms
Authors: Yeganli, Faezeh
Keywords: Electrical and Electronic Engineering
Image processing - Digital techniques
Dual-tree complex wavelet transforms - Q-shift filters
Orthogonal wavelets - Parameterization - Image Denoising
Issue Date: 2010
Publisher: Eastern Mediterranean University (EMU)
Citation: Yeganli, Faezeh. (2010). A Design of Q-shift Filter for Dual-Tree Complex Wavelet Transforms. Thesis (M.S.), Eastern Mediterranean University, Institute of Graduate Studies and Research, Dept. of Electrical and Electronic Engineering, Famagusta: North Cyprus.
Abstract: ABSTRACT: In this work a new method of designing filter for Dual-tree complex wavelet transform is presented. In the new method, the space of orthonormal wavelet filters is defined in terms of some parameters, these parameters are used to design Q-shift filters to have desirable properties including good smoothness and support in [-2p/3, 2p/3]. The constraints in parameterization method lead to wavelets having two vanishing moments. For obtaining the group delay of 1/4 sample period and minimizing the magnitude or energy in stop band [2p/3, p], Kingsbury minimized the energy in this domain. In the proposed method in this work, we minimized the peak magnitude of filters in the stop band. The design approach is illustrated with four examples. The results are compared with Kingsbury’s Q-shift in ana lyticity measures, shift-invariance property and halfsample delay. The designed filters are then used in image denoising. We used the Bivariate shrinkage algorithm for wavelet coefficient modeling and thresholding. Three images (Boat, Baboon, and Cameraman) have been used for test. The experimental results are compared with those obtained using Kingsbury’s Q-shift filters. Keywords: Dual-tree complex wavelet transforms, Q-shift filters, Orthogonal wavelets, Parameterization, Image denoising. …………………………………………………………………………………………………………………………………………………………………………………………………………………… ÖZ: Bu çalismada Ikili agaç kompleksi dalgacik dönüsümü için filtre tasarlamanin yeni bir yöntemi sunulmaktadir. Yeni yöntemde ortonormal dalgacik filtrelerinin alani parametrelerle belirlenmektedir, sonra bu parametreler iyi pürüzsüzlük ve [-2p/3, 2p/3]’de destek de dahil istenen özelliklere sahip Q-shift filtrelerinin tasarlanmasinda kullanilmaktadir. Parametrizasyon yöntemindeki kisitlar dalgaciklarin iki kaybolma hareketine sahip olmasina neden olmaktadir. 1/4 örnek periyodunun grup gecikmesini elde etmek ve [2p/3, p]’de istenmeyen büyüklük veya enerjiyi asgariye indirmek için Kingsbury bu alandaki enerjiyi minimize etmistir. Bu çalismada önerilen yöntemde filtrelerin tepe büyüklügünü söndürme kusaginda minimize ettik. Sekilli örnekler tasarimin yaklasimini göstermektedir ve sonuçlar çözümleyicilik ölçümünde ler, shift-degismezlik özelliginde ve yarim örnek gecikmesinde Kingsbury’nin Q-shift’i ile karsilastirilabilirdir. Tasarlanan filtreler görüntü gürültüsüzlestirmede kullanilmaktadir. Dalgacik katsayi modellemesi ve esiklemesi için iki degiskenli fire algoritmasini kullandik. Test için üç image (Kayik, Babun ve Kameraman) kullanilmistir ve deneysel sonuçlar Kingsbury'nin Q-shift filtrelerinin kullanilmasiyla elde edilenlerle karsilastirilmistir. Anahtar sözcükler: Ikili agaç kompleksi dalgacik dönüsümü, Q-shift filtresi, Ortogonal dalgaciklar, Parametrizasyon, Görüntü gürültüsüzlestirme.
Description: Master of Science in Electrical and Electronic Engineering. Thesis (M.S.)--Eastern Mediterranean University, Faculty of Engineering, Dept. of Electrical and Electronic Engineering, 2010. Supervisor: Prof. Dr. Runyi Yu.
URI: http://hdl.handle.net/11129/57
Appears in Collections:Theses (Master's and Ph.D) – Electrical and Electronic Engineering

Files in This Item:

File Description SizeFormat
FaezehYeganli.pdf1.56 MBAdobe PDFView/Open


This item is protected by original copyright

Recommend this item
View Statistics

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.

 

Valid XHTML 1.0! DSpace Software Copyright © 2002-2010  Duraspace - Feedback