Please wait a minute...
浙江大学学报(工学版)
J4  2012, Vol. 46 Issue (1): 130-135    DOI: 10.3785/j.issn.1008-973X.2012.01.21
信息工程     
曲面表面文字平面化的算法
李响1,严晓浪1,蒋清晓2
1.浙江大学 超大规模集成电路设计研究所,浙江 杭州 310027；
2. 中国残疾人辅助器具中心电子信息技术研究所,浙江 杭州 310012
Algorithm of planarize text on curved surface
LI Xiang1, YAN Xiao-lang1, JIANG Qing-xiao2
1. Institute of VLSI Design, Zhejiang University, Hangzhou 310027, China; 2. China Disabled Persons Assistive
Devices Center Electronic Information Technology Institute, Hangzhou 310012, China
 全文: PDF  HTML
摘要：

针对书籍内缝边缘类型的材料在进行光学字符识别（OCR）时会产生曲面变形的问题,提出将曲面表面的文字平面化的算法.将文字切割为网格,通过网格的尺寸数据,拟合计算材料变形的表示函数,得到变形材料与展平材料之间的映射关系,将变形材料表面还原.提出一种适合快速计算和实现的不需要拟合函数和积分运算的快速算法,经过测试,验证了该算法的适用范围.结果表明,该算法比拟合函数的算法在效率上有很大提高,并且相对原变形材料,识别率大幅上升；该算法不需要复杂的采集设备,可对绝大多数已有的扫描数据进行校正.
Abstract:

In optical character recognition (OCR) works, the edge of book material can produce surface deformation, and these parts can be hard to recognize by software. An algorithm was proposed to solve the problem. The main purpose of the algorithm is to flatten the deformation text information. The mapping information between raw materials and the flattened materials was obtained through cutting the text into grid and fitting calculation of material deformation function of the curvature. Then the flattened information was regenerated. A new algorithm was proposed which can  perform well but doesn’t need too much calculation. The algorithm was tested and proved under different situations. The efficiency was greatly improved, and recognition rate was significantly increased compared with the raw materials. The algorithm does not require complicate device to gather data, and most existing scanned data can be used for correction process.
出版日期: 2012-02-22
:  TN 409  
基金资助:

中华人民共和国科技部科技型中小企业技术创新基金资助项目（07C26213301377）.
通讯作者: 严晓浪，男，教授, 博导.     E-mail: yan@vlsi.zju.edu.cn
作者简介: 李响（1982-），男，博士生，从事电子信息技术在视力障碍领域的应用研究.E-mail: acios@vip.163.com
服务　　
把本文推荐给朋友
加入引用管理器
E-mail Alert
RSS
作者相关文章　　

引用本文:

李响,严晓浪,蒋清晓. 曲面表面文字平面化的算法[J]. J4, 2012, 46(1): 130-135.

LI Xiang, YAN Xiao-lang, JIANG Qing-xiao. Algorithm of planarize text on curved surface. J4, 2012, 46(1): 130-135.

链接本文:

http://www.zjujournals.com/eng/CN/10.3785/j.issn.1008-973X.2012.01.21    或    http://www.zjujournals.com/eng/CN/Y2012/V46/I1/130

［1］ YAMASHITA A, KAWARAGO A, KANEKO T, et al. Shape reconstruction and image restoration for nonflat surfaces of documents with a stereo vision system ［C］∥ 17th International Conference on Pattern Recognition. Cambridge: IEEE, 2004: 482-485．
［2］ WADA T, UKIDA H, MATSUYAMA T. Shape from shading with interreflections under a proximal light source: distortionfree copying of an unfolded book ［J］. International Journal of Computer Vision, 1997, 24(2) : 125-135．
［3］ BROWN M S. Image restoration of arbitrarily warped documents ［J］. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2004, 26(10): 1295-1306．
［4］ GATOS B, PRATIKAKIS I, NTIROGIANNIS K. Segmentation based recovery of arbitrarily warped document image ［C］∥ Proceeding of 9th International Conference on Document Analysis and Recognition. Parana: IEEE, 2007: 989-993．
［5］ CAO H, DING X, LIU C. A cylindrical surface model to rectify the bound document image ［C］∥ Proceeding of 9th IEEE International Conference on Computer Vision. Nice: IEEE, 2003: 228-233．
［6］ TORR P H S, MURRAY D W. The development and comparison of robust methods for estimating the fundamental matrix ［J］. International Journal of Computer Vision, 1997, 24(3): 271-300．
［7］ ZHANG L, TAN C L. Restoring warped document images using shapefromshading and surface interpolation ［C］∥ Proceeding of 18th International Conference on Pattern Recognition. Hong Kong: IEEE, 2006: 642-645．
［8］ LU S, TAN C L. Document flattening through grid modeling and regularization ［C］∥ Proceeding of 18th International Conference on Pattern Recognition. Hong Kong: IEEE, 2006: 971-974．
［9］ LU S, CHEN B M, KO C C. Perspective rectification of document images using fuzzy set and morphological operations ［J］. Image and Vision Computing, 2005, 23(5): 541-553．
No related articles found!