Combined internal and external category-specific image denoising

In this paper, we present a category-specific image denoising algorithm that exploits patch similarity within the input image and between the input image and an external dataset. We rely on standard internal denoising for smooth regions while consulting external images in the same category as the input to denoise textured regions. The external denoising component estimates the latent patches using the statistics, i.e. means and covariance matrices, of external patches, subject to a low-rank constraint. In the final stage, we aggregate results of internal and external denoising using a weighting rule based on the patch SNR measure. Our experimental results on five datasets confirms that the proposed algorithm produces superior results compared with state-of-the-art denoising methods both qualitatively and quantitatively.