Circuital Markov random fields for analog edge detection

A nonlinear analog circuit for coupled image restoration and edge extraction based on Markov Random Fields (MRFs) is presented and analyzed. The circuit is made up of a set of identical linear capacitors and two sets of proper nonlinear resistors. The electric behaviour of the resistive part of the circuit is governed by a generalized potential function whose minima represent the solutions of the MRF problem. The advantages of analogic computations based on MRFs over classical digital approaches are discussed. In particular the real-time MRF-like processing obtained through local interactions between the circuit elements turns out to be the most interesting characteristic of the proposed circuit. Some simulations are described in order to show the performances of the MRF non-linear analog circuit