Optical pattern recognition using wavelet transform

The wavelet transform has acquired growing importance in signal and image processing. We introduce an optical wavelet matched filter that improves the conventional matched filter and can be used for pattern recognition. The wavelet matched filter performs a wavelet transform to enhance significant features of the images and the correlation between two wavelets transforms in a single step. Here we present computer simulation results of this new filter and demonstrate that this new bandpass filter has a much better discrimination capability than the conventional matched filter.