Bilateral Half-Box Image Filtering

Abstract

Because edge-aware filtering has been widely used for image-based automation technologies, developing a practical, fast algorithm for its use is important. This letter proposes a simple and fast computational method for edge-aware image filtering based on bilateral half-box regions. Our filter consists of a weighted average of only eight color sums within each half-box region adjacent to a given pixel, where the tonal weights are similar to those obtained with bilateral filtering. The eight sums are efficiently obtained by a single fast box filter using the relative coordinate relationship between the pixel and each half-box center. We examined the performance of our filter based on comparisons with conventional methods.