Double Weighted RPCA Denoising Algorithm for Color Images
Yisu Zhou, Yuhao Dou · 2018
Image denoising is the process of removing the noise and recovering clean images from noisy images. Lots of algorithms have been proposed in last decades in this field, while most of them are aimed at gray images and ignore the correlation and difference between the RGB channels. In this paper, a double weighted Robust Principal Component Analysis (RPCA) denoising algorithm for color images (C-W2RPCA) is proposed. Superior to previous algorithms, it takes into account the spectral correlation, different noise levels for different channels, as well as the importance degree of image component described by singular values. Experiment shows that the proposed C-W2RPCA outperforms the competing methods.