Title

Analytical Test on Effectiveness of MCDF Operations

Document Type

Conference Proceeding

Publisher

Springer

School

School of Computer and Information Science

RAS ID

3198

Comments

Originally published as: Kong, J., Zhang, B., & Guo, W. (2004). Analytical Test on Effectiveness of MCDF Operations. In Computational Science - ICCS 2004- Proceedings Part II. Lecture Notes in Computer Science, vol 3037. Berlin, Germany: Springer. Original paper available here

Abstract

Modified conjugate directional filtering (MCDF) is a method proposed by Guo and Watson recently for digital data and image processing. By using MCDF, directionally filtered results in conjugate directions can be not only merged into one image that shows the maximum linear features in the two conjugate directions, but also further manipulated by a number of predefined generic 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 results on its effectiveness have been obtained. This has been the major obstacle on the decision whether it is worth developing a usable MCDF system. This paper firstly outlines a FFT-based analytical design for conducting the tests, and then presents the results of applying this analytical design to the analysis of MCDF(add1) operation for an image of digital terrain model in central Australia. The test verifies that the MCDF(add1) operation indeed overcomes the two weaknesses of using the conventional directional filtering in image processing, i.e., separation in presentation of processed results in different directions, and significant loss in low-frequency components. Therefore, the MCDF method is worth for further development.