Skeletonization of Character Based on Wavelet Transform
We propose a novel scheme for computing skeleton of characters using wavelet transform. The development of the method depends on a new wavelet function which is designed specifically for computing the medial axis and corner points of character strokes. Based on some certain desirable properties yielded by this particular wavelet function, wavelet skeleton is defined and computed. By using multiscale corner detection with new wavelet function, a set of modifying techniques are developed to remove the artifacts of primary skeletons. Experimental results show that the proposed algorithm overcomes some undesirable effects and limitations of previous methods.
KeywordsCharacteristic Point Wavelet Transform Corner Point Wavelet Function Medial Axis
Unable to display preview. Download preview PDF.
- 1.Blum, H.: A transformation for extracting new desxriptors of shape. In: Wathen- Dunn, W. (ed.) Models for the Perception of Speech and Visual Form, pp. 362–380. The MIT Press, Massachusetts (1967)Google Scholar
- 2.Brady, M.: Criteria for Representation of Shape. In: Beck, J., Hope, B., Rosenfeld, A. (eds.) Human and Machine Vision, pp. 39–84. Academic Press, New York (1983)Google Scholar
- 7.Leyton, M.: Symmetry, Causality, Mind. The MIT Press, Massachusetts (1988)Google Scholar
- 9.You, X., Tang, Y.Y., Sun, L.: Skeletonization of Ribbon-like Shapes with New Wavelet Function. In: The First International Conference on Machine Learning and Cybernetics, Beijing, China, November 4-5 (2002)Google Scholar