Efficiency Optimization of Trainable Feature Extractors for a Consumer Platform

  • Maurice Peemen
  • Bart Mesman
  • Henk Corporaal
Part of the Lecture Notes in Computer Science book series (LNCS, volume 6915)

Abstract

This paper proposes an algorithmic optimization for the feature extractors of biologically inspired Convolutional Neural Networks (CNNs). CNNs are successfully used for different visual pattern recognition applications such as OCR, face detection and object classification. These applications require complex networks exceeding 100,000 interconnected computational nodes. To reduce the computational complexity a modified algorithm is proposed; real benchmarks show 65 - 83% reduction, with equal or even better recognition accuracy. Exploiting the available parallelism in CNNs is essential to reduce the computational scaling problems. Therefore the modified version of the algorithm is implemented and evaluated on a GPU platform to demonstrate the suitability on a cost effective parallel platform. A speedup of 2.5x with respect to the standard algorithm is achieved.

Keywords

Convolutional Neural Networks Feature Extraction GPU 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. 1.
    Chakradhar, S., Sankaradas, M., Jakkula, V., Cadambi, S.: A dynamically configurable coprocessor for convolutional neural networks. In: ISCA 2010: Proceedings of the 37th Annual International Symposium on Computer Architecture, pp. 247–257. ACM, New York (2010)Google Scholar
  2. 2.
    Farabet, C., Poulet, C., Han, J., LeCun, Y.: Cnp: An fpga-based processor for convolutional networks. In: International Conference on Field Programmable Logic and Applications, FPL 2009, pp. 32–37 (August 2009)Google Scholar
  3. 3.
    Garcia, C., Delakis, M.: Convolutional face finder: A neural architecture for fast and robust face detection. IEEE Trans. Pattern Anal. Mach. Intell. 26(11), 1408–1423 (2004)CrossRefGoogle Scholar
  4. 4.
    Haykin, S.: Neural Networks and Learning Machines, 3rd edn. Prentice Hall, Englewood Cliffs (2008)Google Scholar
  5. 5.
    Hinton, G.E.: Connectionist learning procedures. Artif. Intell. 40(1-3), 185–234 (1989)CrossRefGoogle Scholar
  6. 6.
    LeCun, Y., Boser, B., Denker, J.S., Henderson, D., Howard, R.E., Hubbard, W., Jackel, L.D.: Backpropagation applied to handwritten zip code recognition. Neural Comput. 1(4), 541–551 (1989)CrossRefGoogle Scholar
  7. 7.
    Lecun, Y., Bottou, L., Bengio, Y., Haffner, P.: Gradient-based learning applied to document recognition. Proceedings of the IEEE 86(11), 2278–2324 (1998)CrossRefGoogle Scholar
  8. 8.
    Lecun, Y., Huang, F.J., Bottou, L.: Learning methods for generic object recognition with invariance to pose and lighting. In: In Proceedings of CVPR 2004 (2004)Google Scholar
  9. 9.
    Nixon, M., Aguado, A.S.: Feature Extraction & Image Processing, 2nd edn. Academic Press, London (2008)Google Scholar
  10. 10.
    Nvidia: NVIDIA CUDA C Programming Guide 3.2. NVIDIA Corporation (2010)Google Scholar
  11. 11.
    Peemen, M., Mesman, B., Corporaal, C.: Speed sign detection and recognition by convolutional neural networks. In: Proceedings of the 8th International Automotive Congress, pp. 162–170 (2011)Google Scholar
  12. 12.
    Rumelhart, D.E., Hinton, G.E., Williams, R.J.: Learning internal representations by error propagation. In: Parallel Distributed Processing: Explorations in the Microstructure of Cognition, vol. 1, pp. 318–362. MIT Press, Cambridge (1986)Google Scholar
  13. 13.
    Simard, P., Steinkraus, D., Platt, J.C.: Best practices for convolutional neural networks applied to visual document analysis. In: ICDAR, pp. 958–962 (2003)Google Scholar
  14. 14.
    Szarvas, M., Yoshizawa, A., Yamamoto, M., Ogata, J.: Pedestrian detection with convolutional neural networks. In: Proceedings IEEE Intelligent Vehicles Symposium, Las Vegas, NV, pp. 224–229 (June 2005)Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2011

Authors and Affiliations

  • Maurice Peemen
    • 1
  • Bart Mesman
    • 1
  • Henk Corporaal
    • 1
  1. 1.Eindhoven University of TechnologyThe Netherlands

Personalised recommendations