A wavelet approach to extracting contours of document images
- 1 January 1999
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
Abstract
In this paper, a new method for extracting contours of document images is proposed. It is a wavelet-based method that consists of three steps. First, by orthogonal wavelet decomposition, a document image is decomposed into a set of wavelet approximation and wavelet details. Next, contour extraction is achieved by applying a wavelet reconstruction technique. In the third step, a multiresolution-contours extraction with respect to an iteration wavelet reconstruction is implemented. A novel combination of multiresolution-contours results in clear contours of the document image. This reconstruction procedure is accompanied by an energy distribution estimation from which wavelet decomposition times can be adaptively controlled. Our method has no redundant operation compared to the scheme of wavelet transform. Therefore, the computational time and memory requirement are less than those by wavelet transform.Keywords
This publication has 7 references indexed in Scilit:
- Multiresolution recognition of handwritten numerals with wavelet transform and multilayer cluster neural networkPublished by Institute of Electrical and Electronics Engineers (IEEE) ,2002
- Wavelet correlation signatures for color texture characterizationPattern Recognition, 1999
- Discriminative wavelet shape descriptors for recognition of 2-D patternsPattern Recognition, 1999
- Feature extraction from wavelet coefficients for pattern recognition tasksPublished by Institute of Electrical and Electronics Engineers (IEEE) ,1999
- Wavelet-based multiresolution local tomographyIEEE Transactions on Image Processing, 1997
- Singularity detection and processing with waveletsIEEE Transactions on Information Theory, 1992
- Ten Lectures on WaveletsPublished by Society for Industrial & Applied Mathematics (SIAM) ,1992