Generation of sigmoid probability functions by clipped differential speckle detection

Many low-level image processing problems can be solved with optimisation techniques such as simulated annealing. However, these iterative stochastic techniques are not adapted to real-time applications. We suggest the use of clipped differential detection of speckle with photothyristors as simple optoelectronic processing elements that can be massively integrated to implement parallel simulated annealing. This processing element has optical input/output windows and provides random outputs according to controllable sigmoid probability laws required for simulated annealing. Results show that operating speeds of 100 kHz are achievable with present technology.