Recovering Dynamic Stroke Information of Multi-stroke Handwritten Characters with Complex Patterns

In this paper, we consider the problem of recovering dynamic stroke information from multi-stroke handwritten character images with complex patterns. The characters are assumed to be of a class of characters whose strokes are recursively formulated. By employing the so-called graph theoretic approach, we develop a systematic algorithm for recovering dynamic stroke information from character images in the class of our interest. It is shown that the correctness of algorithm is guaranteed mathematically. Moreover, we show that the time complexity becomes O(n), where n denotes the number of stroke-intersections on characters. Some recovery examples are included.