Since the last special section on wavelet transforms appeared in the September 1992 issue of OpticalEngineering, wavelet transform activity in optics has increased at least an order of magnitude. The last special section was an outcome of the Gordon Research Conference on Holography and Signal Processing. This special section is an outcome of the April 1994 SPIE Wavelet Applications Conference in Orlando, which contained more than 80 papers. From these, a number were selected to be revised and

more »

... ncluded in this special section along with a number of independent submissions. A special emphasis is on data-driven adaptivity in waveform and basis selection to best address particular applications. Several important new theories presented in this special section center on how to widen the flexibility of wavelet transform waveform and basis selection, how to increase the robust nature of the discrete wavelet transform using filter banks, and on the relationship between the discrete adaptive wavelet transform and continuous adaptive wavelet transform. Applications of wavelet transforms are numerous and include classical signal pattern recognition, radar, sonar, speech, sound, medical imaging, and data processing. We believe that this special section will inspire more theories and applications to explore the mathematical freedom of allowing the wavelet transform to choose its own linear transform kernels, to enhance signal-to-noise ratio, and to increase the robustness of the wavelet transform by adaptivity. The adaptivity should match naturally to iterative processors with feedback. We are anticipating another special section in the near future in which wavelet chips for data compression and feature extraction arejoined with neurochips for automatic classification, thus widening the application domain significantly.