Abstract
We designed a variety of k-nearest-neighbor parallel architectures for FPGAs in the form of parameterizable soft IP cores. We show that they can be used to solve large classification problems with thousands of training vectors, or thousands of vector dimensions using a single FPGA, and achieve very high throughput. They can be used to flexibly synthesize architectures that also cover: 1NN classification (vector quantization), multishot queries (with different k), LOOCV cross-validation, and compare favorably to GPU implementations. To the best of our knowledge this is the first attempt to design flexible IP cores for the popular kNN classifier.
- Bennett, K. P. and Mangasarian, O. L. 1992. Robust linear programming discrimination of two linearly inseparable sets. Optimiz. Methods Softw. 1, 23--34.Google Scholar
Cross Ref
- Breast Cancer Dataset. 2010. archive.ics.uci.edu/ml/datasets/Breast+Cancer+Wisconsin+(Diagnostic).Google Scholar
- Breast Cancer Dataset. 1992. http://archive.ics.uci.edu/ml/datasets/Breast+Cancer+Wisconsin+%28 Original%29.Google Scholar
- Cappello, P., Davidson, G., Gersho, A., Koc, C., and Somayazulu, V. 1986. A systolic vector quantization processor for real-time speech coding. In Proceedings of the IEEE International Conference on Acoustic, Speech and Signal Processing. 2143--2146.Google Scholar
- Chen, Y. A., Lin, Y. L., and Chang, L. W. 1992. A systolic algorithm for the k - nearest neighbors problem. IEEE Trans. Comput. 41, 103--108. Google Scholar
Digital Library
- Core Generator Tool. 2008. Xilinx user guide. http://www.xilinx.com/itp/xilinx6/books/docs/cgn/cgn.pdf.Google Scholar
- Duda, R. O., Hart, P. E., and Stork, D. G. 2000. Pattern Classification. Wiley-Interscience. Google Scholar
Digital Library
- Garcia, V., Debreuve, E., and Barlaud, M. 2008. Fast k nearest neighbor search using gpu. In Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops. 1--6.Google Scholar
- Jones, H. D., Powell, A., Bouganis, C. S., and Cheung, Y. K. P. 2010. GPU versus fpga for high productivity computing. In Proceedings of the International Conference on Field Programmable Logic and Applications. 119--124. Google Scholar
Digital Library
- Kestur, S., Davis, D. J., and Williams, O. 2010. BLAS comparison on fpga, cpu and gpu. In Proceedings of the IEEE Annual Symposium on VLSI. 288--293. Google Scholar
Digital Library
- Kung, S. Y. 1988. VLSI Array Processors. Prentice Hall. Google Scholar
Digital Library
- Kurgan, L. A., Cios, K. J., Tadeusiewicz, R., Ogiela, M., and Goodenday, L. S. 2001. Knowledge discovery approach to automated cardiac spect diagnosis. Artif. Intell. Med. 23, 149--169. Google Scholar
Digital Library
- Kwai, D. M. and Parhami, B. 2001. Scalable linear array architecture with data-driven control for ultrahigh speed vector quantization. J. VLSI Signal Process. Syst. 28, 235--243. Google Scholar
Digital Library
- Liu, J., Li, B., and Liang, D. 2007. Design and implementation of fpga-based modified bknn classifier. Int. J. Comput. Sci. Netw. Secur. 7, 67--71.Google Scholar
- Lucas, S. M. 1998. A fast exact parallel implementation of the k-nearest neighbor pattern classifier. In Proceedings of the IEEE World Congress on Computational Intelligence. 1867--1872.Google Scholar
- Mangasarian, O. L., Setiono, R., and Wolberg, W. H. 1990. Pattern recognition via linear programming: Theory and application to medical diagnosis. In Large-Scale Numerical Optimization, T. F. Coleman and Y. Li, Eds., SIAM Publications, Philadelphia, PA, 22--30.Google Scholar
- Manolakos, E. S. and Logaras, E. 2007. High throughput systolic som ip core for fpgas. In Proceedings of the IEEE International Conference on Acoustic, Speech and Signal Processing. II-61-II-64.Google Scholar
- Manolakos, E. S. and Stamoulias, I. 2010. IP-cores design for the knn classifier. In Proceedings of the IEEE International Symposium on Circuits and Systems. 4133--4136.Google Scholar
- Spectf. 2010. Heart dataset. http://archive.ics.uci.edu/ml/datasets/SPECTF+Heart.Google Scholar
- Stokes, L. M. 2007. A brief look at fpgas, gpus and cell processors. Int. Test Eval. Assoc. J. 28, 2, 9--11.Google Scholar
- Tahir, M. A. and Bouridane, A. 2006. An fpga based coprocessor for cancer classification using nearest neighbor classifier. In Proceedings of the IEEE International Conference on Acoustic, Speech and Signal Processing.Google Scholar
- Tahir, M. A., Bouridane, A., Kurugollu, F., and Amira, A. 2005. A novel prostate cancer classification technique using intermediate memory tabu search. EURASIP J. Appl. Signal Process. 14, 2241--2249. Google Scholar
Digital Library
- Thomas, D. B., Howes, L., and Luk, W. 2009. A comparison of cpus, gpus, fpgas, and massively parallel processor arrays for random number generation. In Proceeding of the ACM/SIGDA International Symposium on Field Programmable Gate Arrays. 63--72. Google Scholar
Digital Library
- Wolberg, W. H. and Mangasarian, O. L. 1990. Multisurface method of pattern separation for medical diagnosis applied to breast cytology. Proc. Nat. Acad. Sci. 9193--9196.Google Scholar
- Xilinx Power Tools Tutorial. 2010. http://www.xilinx.com/support/documentation/sw_manuals/xilinx11/ug733.pdf.Google Scholar
- Xst. 2008. User guide 10.1. www.xilinx.com.Google Scholar
- Yeh, Y. J., Li, H. Y., Hwang, W. J., and Fang, C. Y. 2007. FPGA implementation of k-nn classifier based on wavelet transform and partial distance. In Proceedings of the Scandinavian Conference on Image Analysis. 512--521. Google Scholar
Digital Library
Index Terms
Parallel architectures for the kNN classifier -- design of soft IP cores and FPGA implementations
Recommendations
Rough Ensemble Classifier: A Comparative Study
WILF '09: Proceedings of the 8th International Workshop on Fuzzy Logic and ApplicationsCombining the results of a number of individually trained classification systems to obtain a more accurate classifier is a widely used technique in pattern recognition. In this article, we have introduced a rough set based meta classifier (RSM). ...
An FPGA-based architecture for embedded systems performance acceleration applied to Optimum-Path Forest classifier
Classification techniques development constitutes a foundation for machine learning evolution, which has become a major part of the current mainstream of Artificial Intelligence research lines. However, the computational cost associated with these ...
Classification of Gunshots with KNN Classifier
EATIS '18: Proceedings of the Euro American Conference on Telematics and Information SystemsIn this article a system of detection and classification of gunshots is proposed, which consists of using the KNN classifier in the presence and absence of Gaussian additive noise. The results guarantee that the classifier reaches up to 94 % of ...






Comments