. Author manuscript; available in PMC: 2010 Dec 22.

Published in final edited form as: IEEE Trans Pattern Anal Mach Intell. 1999 Oct;21(10):974. doi: 10.1109/34.799905

TABLE 1.

Best Performance for Each Classifier

Optic Flow	Correlation	85.6% ± 3.3
Optic Flow	Smoothed	53.1% ± 4.7
Holistic Spatial Analysis	PCA	79.3% ± 3.9
	LFA	81.1% ± 3.7
	FLD	75.7% ± 4.1
	ICA	95.5% ± 2.0
Local Spatial Analysis	Gaussian Kernel	70.3 ± 4.
	PCA Shift-inv	73.4% ± 4.2
	PCA Shift-var	78.3% ± 3.9
	PCA Jets	72.1% ± 4.2
	Gabor Jets	95.5% ± 2.0
Human Subjects	Naive	77.9% ± 2.5
Human Subjects	Expert	94.1% ± 2.1

PCA: Principal component analysis. LFA: Local feature analysis. FLD: Fisher’s linear discriminant. ICA: Independent component analysis. Shift-inv: Shift-invariant. Shift-var: Shift-variant.