Fast KNN Classification Based On Softcore Cpu And Reconfigurable Hardware

This paper presents a novel architecture for k-nearest neighbor (kNN) classification using field programmable gate array (FPGA). In the architecture, the first k closest vectors in the design set of a kNN classifier for each input vector are first identified by performing the partial distance search (PDS) in the wavelet domain. To implement the PDS in hardware, subspace search, bitplane reduction, multiplecoefficient accumulation and multiple-module computation techniques are employed for the effective reduction of the area complexity and computation latency. The proposed implementation has been embedded in a softcore CPU for physical performance measurement. Experimental results show that the implementation provides a costeffective solution to the FPGA realization of kNN classification systems where both high throughput and low area cost are desired.