A hybrid approach for image half-toning combining simulated annealing and neural networks based techniques: implementation on a zero instruction set computer based neural machine

Simulated annealing based algorithms are a very powerful class of stochastic algorithms for degraded image reconstruction. However, the reconstruction of a degraded image using an iterative stochastic process requires a large number of operations and is still out of real time. On the other hand, the learning and generalization capability of ANN models allows an improvement on classical techniques' limitations. We investigate the parallel implementation of image processing techniques. We present a hybrid approach for image half-toning combining simulated annealing and neural network based techniques. Simulation and experimental results are reported.