Image Representation Using Fuzzy Morphological Wavelet
Chin-Pan Huang · InTech eBooks · 2008
A novel image representation using fuzzy morphological approach has been presented in this paper. Using the fuzzy morphological operators and the minimum and maximum reconstruction we develop the fuzzy morphological interpolation (FMI) algorithm. Based on FMI and the hierarchical pyramid structure, we formulate the analysis and synthesis procedure, similar to those given by wavelet transform. Through using the fuzzy morphological approach, a signal can be efficiently represented with several additional advantages, such as lower computation complexity and easily extend to two dimensions. Furthermore, our representation can be implemented very fast by parallel. We successfully use the fuzzy mathematical morphology approach to extend the work of the Pitas and Venetsanopoulos and of Song and Delp on morphological signal representation. We have applied our representation to image analysis and shape recognition, the experimental results have shown the advantage of using our FMW representation as compare with the WT (Daubechies, 1988) and Fourier descriptor (Persoon & Fu, 1977) methods.