Abstract

Wavelet transform has been widely applied to time-frequency signal analysis, image processing (enhancement, feature extraction, etc.), and target detection. Since wavelet transform is a convolution process between an input and a large number of wavelet bases, the computation load increases nonlinearly with the sizes of the input and the wavelet. Near real-time optical wavelet transform [1-3] could be accomplished by using an optical correlator architecture. The processing speed of optical wavelet transform is independent of the size of the wavelet filter and is only limited to the updating speed of the spatial light modulator.

© 1995 Optical Society of America