Complex Text Location and Segmentation Approach through Region Fusion and Edge Features Extraction Method

Fen Cai; Yan-Min Luo; Pei-Zhong Liu; Hsuan-Ming Feng

doi:10.6180/jase.201812_21(4).0012

Complex Text Location and Segmentation Approach through Region Fusion and Edge Features Extraction Method

Computer Science and Information Engineering

Fen Cai¹, Yan-Min Luo², Pei-Zhong Liu³ and Hsuan-Ming Feng This email address is being protected from spambots. You need JavaScript enabled to view it.⁴

¹College of Mathematics and Computer Science, Quanzhou Normal University, Quanzhou, Fujian 362000, P.R. China
²College of Computer Science & Technology, Huaqiao University, Xiamen 361021, P.R. China
³College of Engineering, Huaqiao University, Quanzhou, Fujian 362021, P.R. China
⁴Department of Computer Science and Information Engineering, National Quemoy University, Kinmen, Taiwan, R.O.C.

Received: November 30, 2017
Accepted: September 11, 2018
Publication Date: December 1, 2018

Download Citation: ||https://doi.org/10.6180/jase.201812_21(4).0012

ABSTRACT

Detecting and locating text information in complex background images has become an important technology for marking the image automatically. The traditional text segmentation method is difficult to obtain the ideal result in the text segmentation of complex background images. According to the features of complex background images, this paper puts forward a new segmentation method by fusing the edge and region features. Firstly, the initial localization of text region is extracted by using stroke filter based on the characteristics of Chinese characters. Secondly, the method of projection histogram analysis is used to remove the conglutination between Chinese characters and background, and the connected region analysis method based on region growing is used to recognize the single or multiple active targets. Finally, the final text segmentation is performed by the improved vertical projection segmentation algorithm which based on the structure of Chinese. The experimental results show that the proposed algorithm has good performance in text location, and can adapt to the characteristics of the complicated background and get accurate character image area comparing with the traditional edge detection method.

Keywords: Edge Feature, Region Feature, Stroke Filter, Text Segmentation

REFERENCES

[1] Wang, Y., H. Zheng, and D. W. Hu (2004) The Technology to Acquire Text in Image and Vedio, Journal of Image and Graphic 9(12), 532538.
[2] Zhong, Y., K. Karu, and A. K. Jain (1995) Locating Text in Complex Color Images, Proceedings of the 3rd International Conference on Document Analysis and Recognition 28(10), 15231535. doi: 10.1109/ ICDAR.1995.598963
[3] Botz-Bornstein, T., and M. M. Mostafa (2017) A Corpus-based Computational Analysis of Philosophical Texts: Comparing Analytic and Continental Philosophy,International Journal of Social and Humanistic Computing 2(3), 230246. doi: 10.1504/IJSHC.2017. 10005737 [4] Zhu, C. J., C. Li, and Z. Xiong (2007) Text Detection and Recognition Technique in Video, Computer Engineering 33(10), 218219.
[5] Keechul, J., K. Kwang, and K. J. Anil (2004) Text Information Extraction in Images and Video: a Survey, Pattern Recognition 37(5), 977997. doi: 10.1016/j. patcog.2003.10.012
[6] Lienhart, R., and A. Wernicke (2002) Localizing and Segmenting Text in Images and Videos, IEEE Transactions on Circuits and Systems for Video Technology 12(4), 256268. doi: 10.1109/76.999203
[7] Mariano, V. Y., and R. Kasturi (2000) Locating Uniform-colored TextinVideo Frames,IEEEProceedings of the 15th International Conference on Pattern Recognition 539542. doi: 10.1109/ICPR.2000.902976
[8] Zhong, Y., H. J. Zhang., and A. K. Jain (2000) Automatic Caption Localization in Compressed Video, IEEE Transactions on Pattern Analysis and Machine Intelligence 22(4), 385392. doi: 10.1109/34.845381
[9] Pan, Y. F., X. Hou, and C. L. Liu (2011) AHybrid Approach to Detect and Localize Texts in Natural Scene Images, IEEE Transactions on Image Processing 20(3), 800 813. doi: 10.1109/TIP.2010.2070803
[10] Zhang, X., and F. C. Sun (2011) Pulse Coupled Neural Network Edge-based Algorithm for Image Text Locating, Tsinghua Science & Technology 1(16), 2230. doi: 10.1016/S1007-0214(11)70004-9
[11] Liu, P. Z., X. Ning, and W. J. Li (2014) Approach for Text Segmentation in Web Image, Computer Engineering and Applications 50(10), 180184.
[12] Shi, Z. G., and L. Q. Gao (2010) New Algorithm for Text Segmentation Based on Stroke Filter, Computer Science 37(1), 287289.
[13] Liu, C. P., G. Liu, C. H. Wang, and Y. Q. Zhu (2013) Design of ImageText Extracting ExperimentBased on Stroke Filter, Research and Exploration in Laboratory 11, 8992.
[14] Liu, Q., C. Jung, and Y. Moon (2006) Text Segmentation Based on Stroke Filter, Proceedings of ACM International Conference on Multimedia, USA, 129 132.
[15] Liu, Q., C. Jung, S. Kim, et al. (2006) Stroke Filter for Text Localization in Video Images, Image Processing, 2006 IEEE International Conference on. IEEE, 14731476. doi: 10.1109/ICIP.2006.312560
[16] Yao, J. L., L. B. Weng, and X. H. Wang (2012) A Text Region Location Method Based on Connected Component, Pattern Recognition and ArtificialIntelligence 25(2), 325331.
[17] Cao, W. (2010) Character Vertical Projection Segmentation Based on Dichotomy, Software Cuide 10, 7172.
[18] Shi, J., and J. Malik (2000) Normalized Cuts and Image Segmentation, IEEE Transactions on Patten Analysis Machine Intelligence 22(8), 888905. doi: 10. 1109/34.868688