发明授权
- 专利标题: Generate-and-test method for column segmentation
- 专利标题(中): 用于列分割的生成和测试方法
-
申请号: US13155011申请日: 2011-06-07
-
公开(公告)号: US08560937B2公开(公告)日: 2013-10-15
- 发明人: Hervé Déjean
- 申请人: Hervé Déjean
- 申请人地址: US CT Norwalk
- 专利权人: Xerox Corporation
- 当前专利权人: Xerox Corporation
- 当前专利权人地址: US CT Norwalk
- 代理机构: Fay Sharpe LLP
- 主分类号: G06F17/27
- IPC分类号: G06F17/27
摘要:
A system, method, and computer program product for segmenting a document are disclosed. The method considers a zone of a document, such as a page frame or other zone which is a predetermined ratio thereof, and while there are remaining elements in the zone, iteratively tests different segmentations of the zone into n candidate columns, and computes a width of a gutter for each n-candidate. Assuming that the gutter width computed meets a threshold test, which may be based on the arrangement of the elements in the columns, and the candidate columns for the n-candidate each contain at least a threshold number of elements, elements are assigned to respective ones of n segmented columns within which they are located. For example, line elements are arranged in blocks of text within the columns, enabling a reading order for sequences of text, such as complete sentences and paragraphs, to be computed.
公开/授权文献
- US20120317470A1 GENERATE-AND-TEST METHOD FOR COLUMN SEGMENTATION 公开/授权日:2012-12-13
信息查询