Parallel Character Reconstruction Expending Compute Unified Device Architecture

Part of the Advances in Intelligent Systems and Computing book series (AISC, volume 177)

Abstract

Neural networks, or the artificial neural networks to be more precise, represents a technology that is rooted in many disciplines: neuroscience, mathematics, statistics, physics, computer science and engineering. Neural network finds applications in such fields as modeling, time series analysis, pattern recognition signal processing and control by virtue of an important property: the ability to learn from input data with or without a teacher .In a biological system, learning involves adjustments to the synaptic connections between neurons same for artificial neural networks (ANNs) works too that has made it applicable to valid applications. Neural Network architecture has the ability to learn for the things and then later on classify the things. Neural Network for Character Recognition is based over Multilayered Architecture having Back-propagation algorithm. First Network is been trained for the alphanumeric handwritten characters and then testing the network with the trained or untrained handwritten characters. We achieved a greater computation enhancement by using modified back- propagation algorithm having an added momentum term, which lowers the training time and speeds the system. The time is more reduced with its parallel implementation using CUDA.

Keywords

CUDA GPU Back-propagation algorithm 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. 1.
    Liu, C.-L.: Normalization-Cooperated Gradient Feature Extraction for Handwritten Character Recognition. Transactions on Pattern Analysis and Machine Intelligence 29(8), 1465–1469 (2007)CrossRefGoogle Scholar
  2. 2.
    Verma, B.K.: Handwritten Hindi Character Recognition Using Multilayer Perceptron and Radial Basis Function Neural Network. In: IEEE Conference on Neural Network, vol. 4, pp. 2111–2115 (1995)Google Scholar
  3. 3.
    Chung, Y.Y., Wong, M.T.: Handwritten Character Recognition by Fourier Descriptors and Neural Network. In: Proceedings of IEEE TENCON 1997. IEEE Region 10 Annual Conference. Speech and Image Technologies for Computing and Telecommunications, vol. 1, pp. 391–394 (1997)Google Scholar
  4. 4.
    Starzyk, J.A., Ansari, N.: Feedforward neural network for handwritten character recognition. In: Proceedings of 1992 IEEE International Symposium on Circuits and Systems, ISCAS 1992, vol. 6, pp. 2884–2887 (1992)Google Scholar
  5. 5.
    Li, Y., Li, J., Meng, L.: Character Recognition Based on Hierarchical RBF Neural Networks. In: Sixth International Conference on Intelligent Systems Design and Applications, ISDA 2006, vol. 1, pp. 127–132 (2006)Google Scholar
  6. 6.
    Smith, S.J., Baurgoin, M.O.: Handwritten character classification using nearest neighbor in large database. IEEE Trans. On Pattern and Machine Intelligence 16(10), 915–919 (1994)CrossRefGoogle Scholar
  7. 7.
    Pal, A., Singh, D.: Handwritten English Character Recognition using Neural Network. International Journal of Computer Science and Communication 1(2), 141–144 (2010)Google Scholar
  8. 8.
    Faaborg, A.J.: Using Neural Networks to Create an Adaptive Character Recognition System, Cornell University, Ithaca, NY (2002)Google Scholar
  9. 9.

Copyright information

© Springer-Verlag Berlin Heidelberg 2013

Authors and Affiliations

  • Anita Pal
    • 1
  • Kamal Kumar Srivastava
    • 1
  • Atul Kumar
    • 1
  1. 1.Department of Computer Science & EngineeringShri Ramswaroop Memorial College of Engineering and ManagementLucknowIndia

Personalised recommendations