Quantitative evaluation on edge preservation in multiresolution images
Ze-Nian Li · International Conference on Image Processing · 1989
Multiresolution image processing and analysis has become popular in recent years. In this approach, image data are represented in a hierarchical manner. The original images are usually placed at the bottom level, Images with reduced resolutions are then generated at successively higher levels. In this paper, a quantitative comparison of five techniques for obtaining multiresolution images is presented. The techniques are hierarchical unweighted averaging (AVG), median filtering based on a 2*2 hierarchy (MED2), median filtering based on a 4*4 hierarchy (MED4), hierarchical selective median filtering (SMED), and alternative sampling (ALT). The purpose of this study is to present and quantitatively evaluate the five image reduction techniques. >