Cuboid filter: simple edge-preserving smoothing

Abstract

In this letter, we propose a new edge-preserving smoothing filter, the cuboid filter. Since many applications for image processing and computer vision are based on the idea of denoising and extracting edges, the edge-preserving smoothing filter can be utilized in many applications, such as high dynamic range image compression, unsharp masking, stylization, and more. For this reason, various edge-preserving smoothing filters, such as the bilateral filter, the guided filter, and the domain transform, have been proposed. In general, these filters face a trade-off relationship between denoising performance and computational speed. The cuboid filter calculates only the mean of 3-dimensional neighboring pixels. The cuboid filter performs satisfactorily in reducing small-amplitude noise and fast filtering. On the basis of qualitative and quantitative evaluations, the effectiveness and the validity of the cuboid filter are shown.