A Novel Image Denoising Method Using Independent Component Analysis and Dual-Tree Complex Wavelet Transform
Zhang Shi, Tingting Tang, Chunli Wu, Ning Xi, Gang Wang · 2010
This paper presents a novel adaptive method of image denoising based on the dual-tree complex wavelet transform (DT-CWT) and independent component analysis (ICA). This method extracts the high-frequency component of the image with the DT-CWT, then combining with the principle of ICA virtual observed noise channel denoise. It does not need a lot of observed image samples, and it is unnecessary to know the detail of the observed image signal type in advance. A single observed image could be denoised by this method adaptively. Moreover, it overcomes the deficiency in wavelet threshold denoising, that is the selective and the quantitive of threshold. The experimental results show that this algorithm denoises the image effectively, improving the SNR (signal noise ratio) & PSNR (peak signal nose ratio) largely and retaining the original image information better.