Modification of Conjugate Directional Filtering: from CDF to MCDF
Document Type
Conference Proceeding
Publisher
ACTA Press
Faculty
Faculty of Computing, Health and Science
School
School of Computer and Information Science
RAS ID
1361
Abstract
We recently reported the trial work using the conjugate directional filtering (CDF) to combine two directional filtered results in conjugate directions into one image that exhibits the maximum linear features in the two conjugate directions. Our further study reveals that the CDF has some weaknesses although it is useful for the enhancement of conjugated linear features. The CDF came initially without consideration of using a weighting system for further data manipulation during operation. CDF-processed image often lacks contrast depth because most background information is removed. In this paper, we present a modification of CDF, named modified conjugate directional filtering (MCDF), in which a weighting system is adopted. This adoption allows not only the linear features in the two conjugate directions to be treated differently, but also some subtle features to be identified. This modification also leads to a proportional combination of a weighted CDF data file and its original file to produce an image on which not only enhanced are the conjugate linear features, but also retained is all the information on the original image. This also makes a MCDF image show 3D effect. A synthetic example is used to test the improvement of the MCDF over the CDF.
Comments
Guo, W., & Watson, A. C. (2002). Modification of conjugate directional filtering: from CDF to MCDF. Proceedings of 2002 IASTED International Conference on Signal Processing, Pattern Recognition and Applications. Calgary, Canada: ACTA Press. Available here