A Fast Self-Organizing Map Algorithm for Handwritten Digit Recognition

  • Yimu Wang
  • Alexander Peyls
  • Yun Pan
  • Luc Claesen
  • Xiaolang Yan
Conference paper
Part of the Lecture Notes in Electrical Engineering book series (LNEE, volume 240)


This paper presents a fast version of the self-organizing map (SOM) algorithm, which simplifies the weight distance calculation, the learning rate function and the neighborhood function by removing complex computations. Simplification accelerates the training process in software simulation and is applied in the field of handwritten digit recognition. According to the evaluation results of the software prototype, a 15–20 % speed-up in the runtime is obtained compared with the conventional SOM. Furthermore, the fast SOM accelerator can recognize over 81 % of handwritten digit test samples correctly, which is slightly worse than the conventional SOM, but much better than other simplified SOM methods.


Neural network Self-organizing map Handwritten digit recognition Simplification 


  1. 1.
    Kohonen T (1990) The self-organizing map. Proc IEEE 78(1):1464–1480CrossRefGoogle Scholar
  2. 2.
    Kohonen T, Kaski S, Lagus K et al (2000) Self organization of a massive document collection. IEEE Trans Neural Netw 11(3):574–585CrossRefGoogle Scholar
  3. 3.
    Silven O, Niskanen M, Kauppinen H (2003) Wood inspection with non-supervised clustering. Mach Vis Appl 3:275–285CrossRefGoogle Scholar
  4. 4.
    Mu-Chun S, Hsiao-Te C (2000) Fast self-organizing feature map algorithm. IEEE Trans Neural Netw 11(3):721–732CrossRefGoogle Scholar
  5. 5.
    Nordström T (1992) Designing parallel computers for self organizing maps. In: Fourth Swedish workshop on computer system architectureGoogle Scholar
  6. 6.
    Lobo VJ, Bandeira N, Moura-Pires F (1998) Distributed Kohonen networks for passive sonar based classification. In: International conference on multisource-multisensor information fusion, Las VegasGoogle Scholar
  7. 7.
    Yaohua Y, Damminda A (2006) Batch implementation of growing self-organizing map. In: International conference on computational intelligence for modelling control and automation, and international conference on intelligent agents, web technologies and internet commerceGoogle Scholar
  8. 8.
    Pena J, Vanegas M (2006) Digital hardware architecture of Kohonen’s self organizing feature maps with exponential neighboring function. In: IEEE international conference on reconfigurable computing and FPGAGoogle Scholar
  9. 9.
    Agundis R, Girones G, Palero C, Carmona D (2008) A mixed hardware/software SOFM training system. Computaciny Sistemas 4:349–356Google Scholar
  10. 10.
    Porrmann M, Witkowski U, Ruckert U (2006) Implementation of self-organizing feature maps in reconfigurable hardware. In: FPGA implementations of neural networks. Springer, Heidelberg, pp 247–269Google Scholar
  11. 11.
    LeCun Y, Cortes C, The MNIST database of handwritten digits.

Copyright information

© Springer Science+Business Media Dordrecht(Outside the USA) 2013

Authors and Affiliations

  • Yimu Wang
    • 1
  • Alexander Peyls
    • 2
  • Yun Pan
    • 1
  • Luc Claesen
    • 2
  • Xiaolang Yan
    • 1
  1. 1.Institute of VLSI DesignZhejiang UniversityHangzhouPeople’s Republic of China
  2. 2.EDMHasselt UniversityDiepenbeekBelgium

Personalised recommendations