EXTRACTION OF CHARACTERS FROM THE QUADTREE ENCODE DOCUMENT IMAGE OF HANGUL

쿼드트리로 구성된 한글 문서 영상에서의 문자추출에 관한 연구

In this paper the method of representing the document image by the quadtree data structure, and extracting each character seperately from the constructed quadtree are described. The document image is represented by a binary encoded quadtree and the segmentation is performed according to the information of each leaf node of the quadtree. Then, each character is extracted by the relation of positions of segments. This method enables to extract characters without examining every pixel in the image and the required storage of document image is decreased.