Skew Detection Algorithm for Form Document Based on Elongate Feature

Autor:	Zhi-guo Jiang, Feng-ying Xie, Lei Wang
Rok vydání:	2007
Předmět:	Pixel Region growing Feature (computer vision) Line (geometry) Skew Function (mathematics) Skew angle Horizontal line test Algorithm Mathematics
Zdroj:	Lecture Notes in Computer Science ISBN: 9783540741954 EMMCVPR
DOI:	10.1007/978-3-540-74198-5_10
Popis:	One new and efficient skew detection algorithm is proposed for form documents according to the feature that the horizontal line has the same skew with the form. This algorithm includes the following steps: Firstly, all horizontal connected regions, including horizontal straight-lines, are extracted from the form document by directional region growing method presented in this paper; Secondly, the optimal line is selected from all horizontal connected regions based on the elongate of connected region; Thirdly, all the pixels belonging to the optimal line are considered to calculate the line parameters with linear least-square theory. The skew angle of the optimal line is just the form skew angle. One elongate function is defined in this paper which described the elongate feature correctly for the band-like connected region. The experiment results show that the form skew angle can be detected accurately, and this skew detection algorithm is fast and robust.
Databáze:	OpenAIRE
Externí odkaz:	https://explore.openaire.eu/search/publication?articleId=doi_________::74ecca0a9592360db4ac0bb58c45ca68 https://doi.org/10.1007/978-3-540-74198-5_10 Zobrazit plný text záznamu