Novel Text Recognition Based on Modified K-Clustering and Hidden Markov Models
- 3 Downloads
Currently, many researchers have paid more attention to identifying scene texts from the image with background interferences. This study aims to develop an App software system with text recognition on smartphones. Otsu edge detection is applied to binarize the image and to find the parameters (i.e. weights) in a K-cluster. The modified K-cluster algorithm is used to detect the text from an image. The noise in complex background is also filtered out. The detected text gradients are evaluated by histogram of gradient. Accordingly, the distribution of the detected text gradients is generated. Finally, the gradient distribution is utilized by hidden Markov models to recognize the text. The experimental results have shown that the proposed approach can successfully outperform other methods.
KeywordsText recognition Edge detection Hidden Markov model Image processing
The authors are very grateful to the anonymous reviewers for their constructive comments which have improved the quality of this paper. Also, this work was supported by the Ministry of Science and Technology, Taiwan, under grant MOST 106- 2221- E-845- 001.
- 3.Judd, T., Ehinger, K., Durand, F., & Torralba, A. (2009). Learning to predict where humans look. In Proceedings of IEEE 12th ICCV (pp. 2106–2113).Google Scholar
- 4.Chen, X., & Yuille, A. (2004). Detecting and reading text in natural scenes. Proceedings of IEEE CVPR,2, 366–373.Google Scholar
- 5.Neumann, L., & Matas, J. (2012). Real-time scene text localization and recognition. In Proceedings of IEEE CVPR (pp. 3538–3545).Google Scholar
- 6.Neuman, L., & Matas, J. (2010). A method for text localization and recognition in real world images. In Proceedings of ACCV (pp. 770–783).Google Scholar
- 7.Odobez, J. M., & Chen, D. (2002). Robust video text segmentation and recognition with multiple hypotheses. In Proceedings of ICIP (pp. 433–436).Google Scholar
- 8.Huang, R., Oba, S., Shivakumara, P., & Uchida, S. (2012). Scene character detection and recognition based on multiple hypotheses framework. In Proceedings of ICPR (pp. 717–720).Google Scholar
- 9.Jetley, S., Behlhe, S., Koppula, V. K., & Nagi, A. (2012). Two-stage hybrid binarization around fringe map based text line segmentation for document images. In Proceedings of ICPR (pp. 343–346).Google Scholar
- 10.Zhang, D., & Chang, S. (2003). A bayesian framework for fusing multiple word knowledge models in videotext recognition. In Proceedings of CVPR (pp. 528–533).Google Scholar
- 11.Lucas, S. M. (2005). Text locating competition results. In Proceedings of third international conference on document analysis and recognition (pp. 80–85).Google Scholar
- 14.Pedro Felipe Felzenszwalb. Introduction to computer vision edge detection [Online]. https://www.classes.cs.uchicago.edu/archive/2008/spring/35040-1/edges.pdf. Accessed 2 June 2017.
- 15.Utrecht University. Chapter 10 segmentation [Online]. http://www.cs.uu.nl/docs/vakken/ibv/reader/chapter10.pdf. Accessed 11 July 2017.
- 17.Wikipedia. Histogram of oriented gradients [Online]. https://en.wiki-pedia.org/wiki/Histogram_of_oriented_gradients. Accessed 11 July 2017.
- 18.Dietterich, Thomas, Bishop, Christopher, Heckerman, David, Jordan, Michael, & Kearns, Michael. (2010). Introduction to machine learning (2nd ed.). London: The MIT Press.Google Scholar
- 22.Davis, R. I. A., Lovell, B. C., & Caelli, T. (2002). Improved estimation of hidden Markov model parameters from multiple observation sequences. Proceedings International Conference on Pattern Recognition,2, 168–171.Google Scholar
- 24.Wang, K., Babenko, B., & Belongie, S. (2011). End-to-end scene text recognition. In Proceedings ICCV (pp. 1457–1464).Google Scholar
- 26.Abbyyfinereader 9.0. http://www.abbyy.com. Accessed 11 July 2017.