Document segmentation is one of the main key element of an OCR system. A recognition system remains inefficient without proper segmentation. With the passing of history, various precious cultural heritage was left behind to tell ancient stories, especially those in the form of written documents. After years of storage, historical document collections encounter serious degradation via staining, tearing, ink seepage, etc. The problem of how to preserve this priceless culture heritage for the next generation has received intense interest from numerous researchers. Historical document digitization is one way to protect such valuable information on historical knowledge and literary arts. Historical documents are digitized through photographing, followed by document segmentation, recognition, preservation, management, and research. Among all the above-mentioned stages, document segmentation is conducted as a first step and the over- all digitization performance of the system heavily depends on the segmentation quality. My project aims at segmenting chinese characters from historical documents for recognition purpose.
