Frame-Based Facial Expression Recognition Using Geometrical Features

Table 2

Confusion matrix of human facial expression recognition for the CK+ database. The first row in each expression represents our results (with SVM classifier). The other row shows the results of Lucey et al., as reported in [13].

	An	Di	Fe	Ha	Sa	Su	Co
An	80.0	6.66	2.22	0.0	11.1	0.00	—
	75.0	7.50	5.00	0.00	5.00	2.50	5.00
Di	6.77	83.0	3.38	0.0	6.77	0.00	—
	5.30	94.7	0.00	0.00	0.00	0.00	0.00
Fe	0.00	16.0	72.0	4.0	8.00	0.00	—
	4.40	0.00	65.2	8.70	0.00	13.0	8.70
Ha	0.00	0.00	0.00	100	0.00	0.00	—
	0.00	0.00	0.00	100	0.00	0.00	0.00
Sa	17.8	10.7	7.14	0.00	64.2	0.00	—
	12.0	4.00	4.00	0.00	68.0	4.00	8.00
Su	0.00	1.25	0.00	0.00	0.00	98.7	—
	0.00	0.00	0.00	0.00	4.00	96.0	0.00