Recovering human-like drawing order from static handwritten images with double-traced lines

Abstract

This paper focuses on an ill-posed problem of recovering a human-like drawing order from static handwritten images with double-traced lines. The problem is analyzed and solved by employing a method based on the graph theoretic approach. Then, a main issue is to obtain the smoothest path of stroke from a graph model corresponding to input image. First, we develop an index on double-traced lines "D-line index" by employing the spline curves. Then, it is shown that the graph is transformed to a semi-Eulerian graph. The restoration problem reduces to maximum weight matching problem and is solved by a probabilistic tabu search. We examine the effectiveness and usefulness by some experimental studies. © 2013 Springer Science+Business Media Dordrecht.