Alphabet Recognition Based on Scaled Conjugate Gradient BP Algorithm

Conference paper
Part of the Lecture Notes in Electrical Engineering book series (LNEE, volume 223)


Artificial neural network is a reflection of brain function to some degree. Neural network has adaptive and self-learning ability and gets features by learning from samples. It can also apply the knowledge which is obtained from learning to the recognition of images, text, and so on. To study alphabet recognition, the Scaled Conjugate gradient BP algorithm is used in this chapter. The simulation results show that, this method can effectively identify the English letters with noise. Compared with the standard BP algorithm, the improved BP algorithm can greatly reduce the training times of the network, and its speed of convergence is much faster.


BP neural network Conjugate gradient algorithm Alphabet recognition 



Institute Level Key Projects Funded by Beijing Institute of Graphic Communication (E-a-2012-31); Funding Project for Academic Human Resources Development in Institutions of Higher Learning under the Jurisdiction of Beijing Municipality (PHR201107145); Scientific Research Common Program of Beijing Municipal Commission of Education of China (KM201210015011).


  1. 1.
    Duan M, He Y (2010) Handwritten alphabet recognition based on LVQ neural network. J Jishou Univ 31(2):41Google Scholar
  2. 2.
    Lv F, Zhu S (2005) Recognition of letters with noise based on hopfield neural network. Comput Inf Technol 12:20Google Scholar
  3. 3.
    Hecht-Nielsen R (1991) Neurocomputing, vol 1. Addison-Wesley Publishing Co, Reading, pp 95–103 Google Scholar
  4. 4.
    Shen Y, Liu L, Ju X (2011) Handwritten character recognition technology based on CMAC neural network. Digit Technol Appl 2:67Google Scholar
  5. 5.
    Wang Z (2011) Vehicle license plate recognition base on improved BP network. Microelectron Comput 28(9):66Google Scholar
  6. 6.
    Xie Z, Shang L, Wang J (2011) A survey combination of rough set and neural. Network 1:1371Google Scholar
  7. 7.
    Yu H, Wu W-Q, Cao L (2009) The improves on the standard BP algorithm and their use. Comput Knowl Technol 5(19):5256Google Scholar
  8. 8.
    Zhu K, Wang Z (2010) Proficient in MATLAB neural network, vol 1, Publishing House of Electronics Industry, Beijing, p 209Google Scholar
  9. 9.
    Cong S, Lu T (2006) Design of three artificial neural networks used on English character recognition. Chin J Sci Instrum 27(6):2242–2244Google Scholar
  10. 10.
    Wan H, Feng X, Lv M (2008) Application of the improved BP algorithm to the recognition of word with noise. Microcomput Inf 24(30):208Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2013

Authors and Affiliations

  1. 1.Beijing Institute of Graphic CommunicationBeijingChina

Personalised recommendations