Abstract

The recognition of signals and images degraded by a linear low-pass degradation operator, even in the absence of any measurement noise, is an ill-posed problem. Based on a new constrained associative memory (CAM) method, the reconstruction (restoration) for an arbitrary binary object from an image, either linearly degraded by a linear shift-invariant (LSI) or shift-variant (LSV) degradation operator, in the presence of strong noise, is achieved. Using an appropriate training set of signals, related ideally to a perfect degradation operator inverse, the CAM method yields a general, in the form of a 2-D array of coefficients, one-step impulsive-type (spiking) inverse filter. Computer simulation results of the reconstruction of 1-D and 2-D signals and images, degraded by LSI and LSV systems, in the presence of strong noise are presented.

© 1988 Optical Society of America