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