Image reinforcement with ALOPEX conducted on connectivity strengths, using moment invariants

ALOPEX is an optimization technique. A cost function is maximized or minimized depending on the problem. In ALOPEX, the cost function is called the response and the many parameters to be optimized are the pixel intensities of patterns. The procedure allows for the parameters to be changed by small amounts in every iteration and the response to be computed. A stochastic element, an added noise, provides for the ability of the process to avoid local maxima or minima. The authors address the issue of the convergence speed using the concept of connectivity strengths of the templates involved, in conjunction with measures of moment invariants