Bivariate shrinkage using undecimated dual-tree complex wavelet transform for image denoising

Article type: Research Article

Authors: Yang, Peng^* | Yang, Guowei

Affiliations: School of Information Engineering, Nanchang Hangkong University, Jiangxi, China

Correspondence: [*] Corresponding author. Peng Yang, School of Information Engineering, Nanchang Hangkong University, Jiangxi, China. Tel./Fax: +86 0791 83953407; E-mail: [email protected].

Abstract: In this paper, we present a bivariate shrinkage method for denoising of images corrupted with additive white Gaussian noise. The proposed method first employs undecimated dual-tree complex wavelet transform on noisy image, which can keep a direct one-to-one relationship between the co-located complex wavelet coefficients at all scales. After that, it estimates the wavelet coefficients by taking into account the parent-child dependency under the non-Gaussian bivariate model, and completes the denoising procedure. Experimental results on test images show that our method can not only eliminate different levels of noise but also obtain fine structures preservation.

Keywords: Image denoising, undecimated dual-tree complex wavelet transform, bivariate shrinkage

DOI: 10.3233/IFS-152037

Journal: Journal of Intelligent & Fuzzy Systems, vol. 30, no. 6, pp. 3111-3122, 2016