We investigate possible performance improvements of coherent optical correlators by using an appropriate filter design. Multidecision strategies are often required in high-level image-processing tasks. For an optical system characterized by a given space–bandwidth product we show that the filter design plays a crucial role in satisfying both system and processing requirements, with respect to the optimization of the encoding capacity. This leads us to the definition of segmented composite filtering, which is discussed in terms of processing performance. This filtering is assessed experimentally in the case of a face-recognition problem.
© 1999 Optical Society of America
(070.2580) Fourier optics and signal processing : Paraxial wave optics
(070.2590) Fourier optics and signal processing : ABCD transforms
(070.4550) Fourier optics and signal processing : Correlators
(070.6760) Fourier optics and signal processing : Talbot and self-imaging effects
Ayman Alfalou, Gilles Keryer, and Jean-Louis de Bougrenet de la Tocnaye, "Optical Implementation of Segmented Composite Filtering," Appl. Opt. 38, 6129-6135 (1999)