Analytical comparison of conventional and MCDF operations in image processing

Abstract

Modified conjugate directional filtering (MCDF), a new method proposed by Guo and Watson in 2002 for digital data and image processing, provides ability in not only integrating directional-filtered results in conjugate directions into one image that shows the maximum linear features in these conjugate directions, but also further manipulating the outcomes using a number of predefined MCDF operations for different purposes. Although a number of cases have been used to test the usefulness of several proposed MCDF operations, and the results are 'visually' better than some conventional methods, however, no quantified analytical comparisons on its effectiveness over conventional methods have been obtained. In this paper, we firstly outline a FFT-based scheme for making analytical comparisons between the results from using conventional and MCDF operations, and then apply this scheme to an aerial photograph of a city view for making comparisons between directional-filtered and MCDF(add2) images. The comparison verifies that the MCDF(add2) operation indeed performs better than conventional directional filtering in image processing in terms of information integration and retention of low-frequency components. © Springer-Verlag Berlin Heidelberg 2004.