Skip to main content
. Author manuscript; available in PMC: 2009 Nov 25.
Published in final edited form as: IEEE Trans Neural Syst Rehabil Eng. 2009 Feb 3;17(2):116–127. doi: 10.1109/TNSRE.2009.2012711

Fig. 9.

Fig. 9

Computational complexity of PCA/EM/Gaussian kernel and the compressed sensing method. (a) Computations per event versus number of events and number of samples per event in the training mode. (b) Computations per event versus number of samples per event and kernel size in the runtime mode. At a sampling rate of 40 KHz and ~1.2–1.5 ms event duration (48–60 samples), the compressed sensing method requires less computations than the PCA/EM/Gaussian kernel method. The number of units is assumed fixed in the training mode for both methods (P = 50).