Using Wavelet Transform and Partial Distance Search to Implement kNN Classifier on FPGA with Multiple Modules

  • Hui-Ya Li
  • Yao-Jung Yeh
  • Wen-Jyi Hwang
Part of the Lecture Notes in Computer Science book series (LNCS, volume 4633)

Abstract

This paper presents a novel algorithm of using wavelet transform and partial distance search (PDS) to realize the kNN classifier on field programmable gate array (FPGA) with multiple modules. The algorithm identifies first k closest vectors in the design set of a kNN classifier for each input vector by performing the PDS in the wavelet domain, and allows concurrent classification of different input vectors for further computation acceleration by employing multiple-module PDS. For the effective reduction of the area complexity and computation latency, we proposed a novel PDS algorithm well-suited for hardware implementation and also employ subspace search, bitplane reduction and multiple-coefficient accumulation techniques. The proposed realization has been embedded in a softcore CPU for physical performance measurements. Experimental results show that the proposed realization not only provides a cost-effective solution to the FPGA implementation of kNN classification systems, but also meets both high throughput and low area cost.

Keywords

FPGA Implementation kNN Classifier Partial Distance Search Pattern Recognition Nonparametric Classification 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. 1.
    Hwang, W.J., Jeng, S.S., Chen, B.Y.: Fast Codeword Search Algorithm Using Wavelet Transform and Partial Distance Search Techniques. Electronic Letters 33, 365–366 (1997)CrossRefGoogle Scholar
  2. 2.
    Hwang, W.J., Wen, K.W.: Fast kNN Classification Algorithm Based on Partial Distance Search. Electronics letters 34, 2062–2063 (1998)CrossRefGoogle Scholar
  3. 3.
    Yeh, Y.J., Li, H.Y., Hwang, W.J., Fang, C.Y.: FPGA Implementation of kNN Classifier Based on Wavelet Transform and Partial Distance Search. In: Proc. SCIA 2007, Aalborg, Denmark, June 10-14, 2007 (2007)Google Scholar
  4. 4.
    Mcnames, J.: Rotated Partial Distance Search for Faster Vector Quantization Encoding. IEEE Signal Processing Letters, 244–246 (2000)Google Scholar
  5. 5.
    Ridella, S., Rovetta, S., Zunino, R.: K-Winner Machines for Pattern Classification. IEEE Trans. Neural Networks 12, 371–385 (2001)CrossRefGoogle Scholar
  6. 6.
    Vetterli, M., Kovacevic, J.: Wavelets and Subband Coding. Prentice Hall, New Jersey (1995)MATHGoogle Scholar
  7. 7.
    Xie, A., Laszlo, C.A., Ward, R.K.: Vector Quantization Technique for Nonparametric Classifier Design. IEEE Trans. Pattern Anal. Machine Intell. 15, 1326–1330 (1993)CrossRefGoogle Scholar
  8. 8.
    Stratix Device Handbook (2005), http://www.altera.com/literature/lit-stx.jsp
  9. 9.
    Custom Instructions for NIOS Embedded Processors, Application Notes 188 (2002), http://www.altera.com/literature/lit-nio.jsp

Copyright information

© Springer-Verlag Berlin Heidelberg 2007

Authors and Affiliations

  • Hui-Ya Li
    • 1
  • Yao-Jung Yeh
    • 1
  • Wen-Jyi Hwang
    • 1
  1. 1.Graduate Institute of Computer Science and Information Engineering, National Taiwan Normal University, Taipei, 117Taiwan

Personalised recommendations