A method for character string extraction using local and global segment crowdedness

We propose a new method for extracting character strings which have various directions and sizes. In our method, we regard a character string as a rectangular region crowded with short segments. Then two kinds of the segment crowdedness (local segment crowdedness and global segment crowdedness) are introduced to extract the rectangular regions (i.e. strings). The method was applied to many images which involve strings in various directions and sizes. The result shows that almost all strings were correctly extracted