Stochastic artificial retinas: algorithm, optoelectronic circuits, and implementation

An analogy can be established between image processing and statistical mech anics. Many early- and intermediate-vision tasks such as restoration, image segmentation, and motion detection can be formulated as optimization probl ems that consist in finding the ground states of an energy function. This a pproach yields excellent results, but, once it is implemented in convention al sequential workstations, the computational loads are too extensive for p ractical purposes, and even fast suboptimal optimization approaches are not sufficient. We elaborate on dedicated massively-parallel integrated circui ts, called stochastic artificial retinas, that minimize the energy function at a video rate. We consider several components of these artificial retina s: stochastic algorithms for restoration tasks in the presence of discontin uities, dedicated optoelectronic hardware to implement thermal motion by ph otodetection of speckles, and hybrid architectures that combine optoelectro nic, asynchronous-analog, and clocked-digital circuits. (C) 2001 Optical So ciety of America.