全部 标题 作者
关键词 摘要

OALib Journal期刊
ISSN: 2333-9721
费用:99美元

查看量下载量

相关文章

更多...

基于投影轮廓的文本图像倾斜检测

DOI: 10.11834/jig.20150104

Keywords: 倾斜检测,投影直方图,方差,RADON变换,降采样

Full-Text   Cite this paper   Add to My Lib

Abstract:

目的在光学字符识别中,文本图像经常会出现一定角度的倾斜。为将倾斜的文本图像校正,以便于字符识别中的后续处理,快速准确地检测倾斜文本图像的倾角是非常重要的。方法对基于投影轮廓的算法进行改进,提出了一种两级投影直方图方差的算法(TPHV)。首先在预定的角度范围内以一定角度步长对选定的图像区域做多方向投影,以获取投影直方图;然后计算各角度投影直方图的均方差,求出所有投影直方图方差的最大差分,将对应的投影角度作为倾角的粗略估值,最后以粗略估值为中心,以第1次投影步长为半径的角度范围内,再次以给定的检测精度为步长进行投影,重复第1次投影的工作,求出投影直方图方差的最大值,以对应的角度作为图像倾角的检测值。结果该算法能够处理各种复杂的文本图像;对于诸如2480×3508像素的较大图像,可在200ms左右的时间内完成倾角的检测;可检测的倾角范围不受限制;对相关网站提供的5组共500幅测试图像检测误差绝对值均值不超过0.5°,最大值不超过0.7°,检测误差的方差不超过0.1。结论实验结果表明,该算法具有明显优势:速度快,倾斜角度检测精度高,误差集中,检测范围大,对噪声不敏感,具有广泛的适用性,适合于复杂的排版方式。

References

[1]  Mehdi F, Nicolas B, Salvatore T. A robust skew detection method based on maximum gradient difference and R-signature[C]//Proceedings of the 18th IEEE International Conference on Image Processing. Washington DC: IEEE, 2011: 2617-2620.
[2]  Li S T, Shen Q H, Sun J. Skew detection using wavelet decomposition and projection profile analysis [J]. Pattern Recognition Letters, 2007, 28: 555-562.[DOI: 10.1016/j.patrec.2006.10.002]
[3]  Javad S, Mohamed C. A new approach for skew correction of documents based on particle swarm optimization[C]//Procee-dings of the 10th International Conference on Document Analysis and Recognition. Barcelona, Spain: IEEE, 2009:1060-1070.
[4]  Lu Y, Tan C L. Improved nearest neighbor based approach to accurate document skew estimation [C]//Proceedings of 7th International Conference on Document Analysis and Recognition. Washington DC: IEEE, 2003: 503-507. [DOI: 10.1109/ICDAR.2003.1227716]
[5]  Michael M, Nikos N, Nikos P. An adaptive technique for global and local skew correction in color documents [J]. Expert Systems with Applications, 2010, 37: 6832-6843. [DOI: 10.1016/j.eswa.2010.03.041]
[6]  Sun N, Lu Z W. An improved method for skew detection of Chinese document images [J]. Computer Simulation, 2006, 23(9): 184-187. [孙楠,刘志文.一种改进的中文文档图像倾斜检测方法[J]. 计算机仿真, 2006, 23(9): 184-187.][DOI:10. 3969/j.issn.1006-9348.2006.09.048]
[7]  Yu B, Jain A K. A robust and fast skew detection algorithm for generic documents [J]. Pattern Recognition,1996,29(10):1599-1629. [DOI: 10.1016/0031-3203(96)00020-9]
[8]  Chandan S, Nitin B, Amandeep K. Hough transform based fast skew detection and accurate skew correction methods [J]. Pattern Recognition, 2008, 41: 3528-3546. [DOI:10.1016/j.patcog.2008.06.002]
[9]  Gao Y P, Li Y M, Hu Z Y, et al. Improved correction method of tilt documentation based on the Hough transform [J]. Journal of Zhejiang University of Technology, 2013, 41(1): 106-109. [高宇鹏, 李益明, 胡众义, 等. 基于Hough变换倾斜文档校正的改进方法[J]. 浙江工业大学学报,2013, 41(1): 106-109.][DOI:10. 3969/j. issn.1006-4303.2013.01.024]
[10]  Deepak K, Dalwinder S. A review of scanned document skew detection and correction techniques [J]. International Journal of Computer Science and Technology, 2012, 3, (2):251-254.
[11]  Zhang J L,Wang X C, Liu J. Document image skew detection based on mathematical morphology [J]. Journal of Computer Applications, 2008, 28: 213-214. [张吉玲,王希常,刘江.基于数学形态学的文档图像倾斜校正算法[J]. 计算机应用, 2008, 28: 213-214.]
[12]  Chou C H, Chu S Y, Chang F. Estimation of skew angles for scanned documents based on piecewise covering by parallelograms [J]. Pattern Recognition, 2007, 40: 443-455.
[13]  Xia B Y, Tong H C. A document image skew detection method based on texture gradation [J]. Computer Simulation, 2009, 26(3): 240-243. [夏波涌, 童悍操. 基于纹理梯度的文档图像的倾斜校正方法[J].计算机仿真,2009, 26(3): 240-243.][DOI: 10.3969/j.issn.1006-9348.2009.03.063]

Full-Text

Contact Us

service@oalib.com

QQ:3279437679

WhatsApp +8615387084133