Image demosaicking by nonlocal adaptive thresholding

Single sensor digital color cameras capture only one of the three primary colors at each pixel and full color image is formed by interpolating all other missing color samples at each pixel. This process is color demosaicking. Most demosaicking algorithms assume the existence of high local spectral redundancy in estimating the missing color samples. However, for images with sharp color transitions and high color saturation, such an assumption may be invalid and visually unpleasant demosaicking errors will occur. In this paper the image non-local redundancy is used to improve the local color reproduction result. For that First, multiple local directional estimates of a missing color sample are computed and fused according to local gradients. Then nonlocal pixels similar to the estimated pixel are searched to enhance the local estimate. For that an adaptive thresholding method rather than the commonly used nonlocal means filtering is implemented to improve the local estimate. Results demonstrate that the local directional interpolation and nonlocal adaptive thresholding (LDI-NAT) method outperforms than other CDM methods in reconstructing the edges and reducing color interpolation artifacts, leading to higher visual quality of reproduced color images.