Active Contours With Stochastic Fronts and Mechanical Topology Optimization

Abstract

Active contours with stochastic fronts (ACSF) is developed and investigated in this article to address the dependency of the existing algorithms of topology optimization on the initial guesses. The promising results of ACSF confirms that the use of this approach leads to higher chance of escaping from local solutions compared to the classic level set method. ACSF as a special case of the stochastic active contours (SAC), has a simplified structure that makes its implementation easier, and at the same time it has a rigorous mathematical proof of convergence. Although propitious, there is still a slight chance of trapping scenarios for ACSF that is observed in the presented results.