A Unified Factors Analysis Framework for Discriminative Feature Extraction and Object Recognition

Table 1

Classification accuracies of LDA and FA-LDA on the artificial database.

3 classes		4 classes		5 classes
FA-LDA	LDA	FA-LDA	LDA	FA-LDA	LDA
83.5 ± 7.45	62 ± 8.34	75.5 ± 5.6	38.54 ± 10.21	78.26 ± 5.4	45.26 ± 7.85
81 ± 9.01	56.7 ± 7.45	73 ± 6.68	45.85 ± 7.37	75.66 ± 6.28	46.15 ± 8.21
83.5 ± 9.07	59.5 ± 10.53	73.5 ± 5.62	46.33 ± 5.36	75.5 ± 7.25	39.54 ± 9.25
80.5 ± 8.67	57 ± 9.57	76.67 ± 5.6	42.64 ± 7.6	79.45 ± 5.5	48.26 ± 6.51
81 ± 9.01	53.5 ± 12.45	78.56 ± 5.21	43.85 ± 7.49	76.5 ± 7.50	44.58 ± 9.21
81.5 ± 6.26	61.5 ± 9.57	78.85 ± 5.34	40.28 ± 6.35	70.57 ± 9.14	39.83 ± 5.68
84.5 ± 6.58	60 ± 6.7	80.21 ± 8.63	42.84 ± 5.49	75.64 ± 5.60	46.25 ± 7.35
82.5 ± 8.08	59.5 ± 11.56	78.26 ± 5.23	41.45 ± 8.16	76.42 ± 5.60	42.35 ± 5.36
81 ± 8.45	64.5 ± 6.87	75.95 ± 6.72	43.58 ± 6.53	77.38 ± 6.58	46.25 ± 8.24