Table 1.

MSE, $E_{\mathrm{p}}$, $R^2$, and AUC values for dataset I from different motion reduction algorithms and no correction. The processing time is calculated for 14 channels (6.6 min of recording with 50-Hz sampling rate) on a 3.4-Hz CPU running windows 7. The bold values are the best values among all algorithms. The proposed method is highlighted in italics.

Dataset I	$\mathrm{MSE}\times {10}^4$	$E_{\mathrm{p}}\times {10}^4$	$R^2$	AUC	Processing time (s)
No correction	$2.26\pm 0.38$	$12.58\pm 2.34$	$0.68\pm 0.02$	$0.87\pm 0.03$	—
Spline	$1.59\pm 0.45$	$8.75\pm 2.71$	$0.73\pm 0.03$	$0.89\pm 0.02$	0.8
Wavelet	$1.22\pm 0.28$	$6.96\pm 1.71$	$0.73\pm 0.03$	$0.88\pm 0.03$	135
CBSI	$2.23\pm 0.40$	$13.48\pm 2.60$	$0.77\pm 0.02$	$\mathbf{0.91}\pm \mathbf{0.03}$	0.2
tPCA	$1.14\pm 0.27$	$7.30\pm 1.98$	$0.71\pm 0.03$	$0.85\pm 0.03$	1.8
tPCA-SG	$0.89\pm 0.21$	$5.50\pm 1.35$	$0.74\pm 0.02$	$0.85\pm 0.03$	14
SG	$1.50\pm 0.31$	$9.32\pm 2.28$	$0.74\pm 0.03$	$0.87\pm 0.02$	0.08
Spline-SG	0.65 ± 0.17	4.20 ± 1.18	0.79 ± 0.02	0.89 ± 0.02	6
Rloess	$1.30\pm 0.33$	$8.19\pm 2.27$	$0.79\pm 0.02$	$0.90\pm 0.02$	764
Spline-Rloess	$\mathbf{0.60}\pm \mathbf{0.16}$	$\mathbf{3.90}\pm \mathbf{1.13}$	$\mathbf{0.80}\pm \mathbf{0.02}$	$0.89\pm 0.02$	767