Table 5.

Cross-conditional classification performance using different sensor domain features

Features	Classifier	Classification performance (%)
		Necker cube (NC)			Stroboscopic alternative motion (SAM)			Structure from motion (SFM)
		Accuracy	Sensitivity	Specificity	Accuracy	Sensitivity	Specificity	Accuracy	Sensitivity	Specificity
Global	SVM	Training			78.97	84.11	73.83	82.07	87.88	76.26
		71.05	82.04	60.06	Training			79.80	86.87	72.73
		70.28	77.40	63.16	74.77	80.37	69.16	Training
	ANN	Training			78.32	77.48	79.16	81.30	82.47	80.13
		69.88	73.31	66.44	Training			80.09	82.22	77.95
		71.38	73.51	69.24	77.34	78.64	76.03	Training
Left hemispheric	SVM	Training			78.50	82.24	74.77	81.82	88.38	75.25
		71.05	81.11	60.99	Training			82.83	91.41	74.24
		71.21	78.95	63.47	74.30	77.57	71.03	Training
	ANN	Training			75.68	77.99	73.36	81.30	84.42	78.18
		71.69	74.72	68.65	Training			81.11	83.31	78.91
		70.65	74.57	66.73	74.39	75.98	72.80	Training
Right hemispheric	SVM	Training			71.96	78.50	65.42	75.25	84.34	66.16
		64.55	72.45	56.66	Training			74.24	80.81	67.68
		69.81	79.26	60.37	73.36	77.57	69.16	Training
	ANN	Training			71.45	73.46	69.44	75.96	81.06	70.86
		64.24	67.72	60.76	Training			73.59	76.31	70.86
		67.38	69.41	65.34	72.64	72.85	72.43	Training
Local	SVM	Training			78.01	80.61	75.42	80.27	84.67	75.86
		67.86	66.30	69.43	Training			77.32	75.53	79.12
		71.54	73.00	70.08	74.91	73.79	76.03	Training
	ANN	Training			77.06	76.96	77.15	78.12	80.28	75.96
		69.08	71.07	67.09	Training			74.76	75.03	74.49
		72.43	74.04	70.82	76.05	77.06	75.05	Training