Side Window Weighted Median Image Filtering
Lianqing Shu, Hui Du · 2020
Window-based processing is one of the most common operations in image filtering. Traditional median filters, involving unweighted and weighted filters, almost always align the center of the neighboring window with the pixel being processed. In this paper, we present an improved version of the popular weighted median filter. We introduce side window weighted median filtering technique, which treats each pixel as a potential edge. Instead of aligning the center of the filtering window, the proposed algorithm aligns the sides or corners of the window with the pixels being processed. Experiments show that the proposed side window weighted median filter owns better edge preserving capabilities than traditional weighted median filter.