Based on the observation that human visual system connects broken edge segments according to long-range pattern contents, the authors here describe a general methodology for pattern reconstruction from scanning electron microscope (SEM) based on intuitive conjectures of human intelligence. In particular, we formulate the problem of pattern reconstruction as a problem of minimizing an objective function, which includes the information of pattern complexity and template matching. This approach of objective function minimization is also consistent with the famous "Occam's razor" principle. They then describe a potential objective function formulation, introduce implementations with acceptable computation complexity, and demonstrate reconstruction results on edge ridge signals extracted from real SEM images. This methodology can greatly improve the robustness of SEM image processing and has the potential to be applied to many other fields, such as computer vision and robust human voice detection as well as inspection of nanoscale structures in the SEM. (c) 2006 American Vacuum Society.