Extended-dimensional anisotropic diffusion using diffusion paths on inter-color planes for noise reduction

This paper presents an anisotropic diffusion (AD)-based noise reduction that extends the diffusion dimensions of a typical AD by producing diffusion paths on inter-color planes. To properly utilize an inter-color correlation for the AD-based noise reduction, inter-color planes from different color planes are predicted by adjusting their local mean values. Then, diffusion path-based kernels (DPKs) for the current color plane and predicted inter-color planes (PIPs) are generated to transform the iterative AD into a single-pass smoothing, which can avoid iterative region analysis. Simultaneously, a regionally and directionally varying diffusion threshold is adopted for the current color plane to preserve image details and to improve the quality of noise elimination near strong edges. For the PIPs, diffusion thresholds are regionally adjusted depending on local correlations between the current color plane and each of the PIPs to optimize the performance of noise reduction obtained from the extended diffusion dimension. Lastly, DPK-based filtering is performed in the current color plane and PIPs by using selected diffusion thresholds for the noise reduction. The experimental results demonstrate that the proposed method successfully improves the quality of denoising by greatly increasing the peak signal-to-noise ratio and structural similarity indexes by up to 4.921 dB and 0.090, respectively, compared with benchmark methods. In addition, the proposed method effectively reduces the computational complexity by avoiding the use of an expensive region analysis.