Wavelet orthonormal decompositions for extracting features in pattern recognition

In this paper, a novel approach based on the wavelet orthonormal decomposit ion is presented to extract features in pattern recognition. The proposed a pproach first reduces the dimensionality of a two-dimensional pattern, and thereafter performs wavelet transform on the derived one-dimensional patter n to generate a set of wavelet transform subpatterns, namely, several uncor related functions. Based on these functions, new features are readily compu ted to represent the original two-dimensional pattern. As an application, e xperiments were conducted using a set of printed characters with varying or ientations and fonts. The results obtained from these experiments have cons istently shown that the proposed feature vectors can yield an excellent cla ssification rate in pattern recognition.