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...

采用信号匹配的支票图像二值化

DOI: 10.13190/jbupt.200602.46.zhanghg, PP. 46-49

Keywords: 投影信号,信号匹配,打印支票日期区域,二值化

Full-Text   Cite this paper   Add to My Lib

Abstract:

提出了一种基于信号匹配的低信噪比图像的信号提取方法,解决类似支票日期域这种既有复杂背景,又有印章噪声干扰的图像二值化问题.这种二值化方法完全不同于传统的方法,它从信号处理的角度出发,利用了部分先验知识和理想状态下的投影轮廓信号,再通过用不同阈值分割的投影信号与之匹配,匹配度最大时的阈值即为图像分割的最佳阈值.本方法比其他的传统的二值化方法更具有自适应性和鲁棒性,通过大量的实验数据,以及与常用的其他算法进行的比较得到了验证.

References

[1]  Kittler J, Illingworth J. Minimum error thresholding
[2]  [J] Pattern Recognition, 1986, 19(1): 41-47.
[3]  张闯,吴铭,郭军. 非限制自由手写粘连数字串数字个数的判断
[4]  [J] 北京邮电大学学报, 2004, 27(3): 63-67. Zhang Chuang, Wu Ming, Guo Jun. Algorithm of the length estimation of unconstrained handwritten connected numeral strings
[5]  [J] Journal of Beijing University of Posts and Telecommunications, 2004, 27(3): 63-67.
[6]  Parker J R. Gray level thresholding in badly illuminated image
[7]  [J] IEEE Trans Pattern Analysis and Machine Intelligence, 1991, 13(8): 813-819.
[8]  Liu K, Suen C Y, Cheriet M, et al. Automatic extraction of baselines and data from check image
[9]  [M] Singapore:World Scientific Publishing Co Pte Ltd, 1997: 213-233.
[10]  Dimauro G, Impedovo S, Pirlo G, et al. Automatic bankcheck processing: a new engineered system
[11]  [M] Singapore:World Scientific Publishing Co Pte Ltd, 1997: 5-42.
[12]  Okada M, Shridhar M. Extraction of user entered components from personal bankchecks using morphological subtraction, automatic bankcheck processing
[13]  [M] Singapore:World Scientific Publishing Co Pte Ltd, 1997:237-253.
[14]  Dawoud A, Kamel M. Iterative multi-model sub-image binarization for handwritten character segmentation
[15]  [J] IEEE Trans Image Processing, 2004, 13(9): 1223-1233.
[16]  刘刚, 魏锋, 吴铭, 等. 基于LDP算法的手写数字串切分
[17]  [J] 北京邮电大学学报, 2003, 26(1): 14-18. Liu Gang, Wei Feng, Wu Ming, et al. A segmentation method of cursive handwritten digit string based on limited dynamic programming
[18]  [J] Journal of Beijing University of Posts and Telecommunications, 2003, 26(1):14-18.

Full-Text

Contact Us

service@oalib.com

QQ:3279437679

WhatsApp +8615387084133