Regeneration of strokes from Chinese calligraphic images

Abstract

In ancient China, people wrote with brush to form Chinese calligraphy. With the rapid development of information technology, people can simply scan this Chinese calligraphy and store it in the form of images. In this project, we want to analysis this Chinese calligraphic images in order to extract the underlying stokes, however, this is more than just an image segmentation problem as the strokes may overlap with one another and the stroke width may not be constant. Besides, we will also determine the number of strokes, location of the strokes, drawing order between all stokes and the drawing direction of each stoke. Finally, we will have an animation on the Chinese calligraphy image of how it is written.