完整後設資料紀錄
DC 欄位語言
dc.contributor.authorTseng, YHen_US
dc.contributor.authorLee, HJen_US
dc.date.accessioned2014-12-08T15:27:16Z-
dc.date.available2014-12-08T15:27:16Z-
dc.date.issued1998en_US
dc.identifier.isbn0-8186-8512-3en_US
dc.identifier.issn1051-4651en_US
dc.identifier.urihttp://hdl.handle.net/11536/19509-
dc.description.abstractCharacters sometimes overlap with non-textual lines in form documents and these interfered-characters would generally be recognized with poor accuracy. In this paper we propose a two-step interfering-line removing method. Positions and orientations of interfering-lines are first detected by the Hough transform. Interfering-line widths are then determined from projection histograms. An ambiguous area is defined to bound an interfering-line. Black runs in the ambiguous are classified into four types and grouped into run-groups. The directions of hidden character strokes in each run-groups are predicted. Black pixels located in these hidden strokes are regarded as character pixels and the other black pixels are considered as interfering pixels, which will be removed then. Most OCR engines are trained by non-interfered sample characters. In order to recognize interfered-characters, we adjust the feature values by assigning a stability value to each of sub-regions. In our collected 1820 interfered handwritten Chinese characters, the recognition accuracy was 24.02% for interfered-characters and 89.91% for characters after removing interfering-lines.en_US
dc.language.isoen_USen_US
dc.titleInterfered-character recognition by removing interfering-lines and adjusting feature weightsen_US
dc.typeProceedings Paperen_US
dc.identifier.journalFOURTEENTH INTERNATIONAL CONFERENCE ON PATTERN RECOGNITION, VOLS 1 AND 2en_US
dc.citation.spage1865en_US
dc.citation.epage1867en_US
dc.contributor.department資訊工程學系zh_TW
dc.contributor.departmentDepartment of Computer Scienceen_US
dc.identifier.wosnumberWOS:000075882100495-
顯示於類別:會議論文