Target matching via edge features has been a hot research topic. Potential-based shape matching, as a relatively new branch, is inspired by the interactions among particles in physics. In more detail, pixels along the target contour and along the text image contours are regarded as two groups of atoms. With appropriate geometric transformations, both atom groups can lead to the strongest attraction, which implies the optimal match. The search process for satisfactory geometric transformations can be considered as a numerical optimization problem, which is handled by stochastic fractal search (SFS) algorithm. Comparative simulations are conducted to investigate the performances of various similarity criteria and the performance of SFS algorithm. In addition, capability to recognize more than one correct target in the test image is emphasized. Theoretical analyses preliminarily indicate that our atomic potential matching model is relevant to an existing technique named lateral inhibition in image enhancement. The goal of this study is to complete the groundwork for further research into highly efficient matching performance.