Research on a Characteristic Extraction Algorithm Based on Analog Space-Time Process for Off-Line Handwritten Chinese Characters

Abstract:

On the absence of space-time information, it is difficult to extract the character stroke feature from the off-line handwritten Chinese character image. A feature extraction algorithm is proposed based on analog space-time process by the process neural network. The handwritten Chinese character image is transformed into geometric shape by different types, different numbers, different locations, different orders and different structures of Chinese character strokes. By extracting fault-tolerant features of the five kinds of the off-line handwritten Chinese characters, the data-knowledge table of features is constructed. The parameters of process neural networks are optimized by Particle Swarm optimization (PSO). The handwritten Chinese characters are used to carry out simulation experiment in SCUT-IRAC-HCCLIB. The experiment results show that the algorithm exhibits a strong ability of cognizing handwritten Chinese characters.

• Periodical: Advanced Materials Research (Volumes 433-440)
• Pages: 3649-3655
• DOI: https://doi.org/10.4028/www.scientific.net/AMR.433-440.3649
• Online since: January 2012
• Authors: Cheng Hui Zhu, Wen Jun Xu, Jian Ping Wang, Xiao Bing Xu
• Keywords: Bionic Pattern Recognition, Fault-Tolerant Representation, Particle Swarm, Process Neural Network

References

[1] T Hildebrandt, W Liu, Optical recognition of handwritten Chinese characters advances since 1980, Pattern Recognition, 1993, 26(2), pp.205-225.

DOI: 10.1016/0031-3203(93)90030-z

[2] Sargur N. Srihari, Xuan-shen Yang, Gregory R. Ball, Off-line Chinese handwriting recognition: an assessment of current technology, Front Computer Science of China, 2007, 1(2), pp.137-155.

DOI: 10.1007/s11704-007-0015-2

[3] Zhu Xiao-ping, Gu Hong-bin, Contemporary research Chinese characters and words recognition, Psychological Science, 1992, 15(1), pp.40-45.

[4] Wang Jian-ping, Zhao Li-xin, Wang Jin-ling, A study of Chinese characters code of bearable mistakes method for off-line handwritten Chinese characters recognition, Journal of Image and Graphics, 2007, 12(12), pp.2169-2178.

[5] Wang Jian-ping, Li Wei-tao, Wang Jin-ling, Luo Guo-jun, Research on Handwritten Chinese characters recognition based on bionic pattern recognition, Journal of Image and Graphics, 2007, 12(7), pp.1261-1269.

[6] Kim H J, Kim K H, Kim S K, et al. On-line recognition of handwritten Chinese characters based on hidden Markov models, " Pattern Recognition, 1997, 30(9), pp.1489-1500.

DOI: 10.1016/s0031-3203(96)00161-6

[7] He Xin-gui, Xu Shao-huo, Process neural network, Science Press, (2007).

[8] Bian Zhao-qi, Zhang Xue-gong, et al, Pattern Recognition, Beijing: Tsinghua University Press, (2000).

[9] Dong Jian-xiong, Adam Krzyzak, Ching Y. Suen, An improved handwritten Chinese character recognition system using support vector machine, Pattern Recognition Lettersm2005, 26(12), pp.1849-1856.

DOI: 10.1016/j.patrec.2005.03.006

[10] Tian Yu-bo, Hybrid neural network technology, Science Press, (2009).