Multipose Face Recognition-Based Combined Adaptive Deep Learning Vector Quantization

Table 3

The GAR (%) of the frontal view (V1), and multipose (V2, V3) for combined ADLVQ classifier.

Multipose face images (V1, V2, V3)	SDUMLA-HMT			CASIA-V5
	V1	V2	V3	V1	V2	V3
ADLVQ-1[PS1]	96.87	90.56	90.00	95.87	90.00	89.00
ADLVQ-2[PS2]	97.05	91.54	90.56	95.92	90.20	89.64
ADLVQ-3[PS3]	97.85	91.65	90.12	95.00	91.38	91.21
ADLVQ-4[PS4]	97.56	92.00	90.87	95.35	92.00	89.66
ADLVQ-5[PS5]	97.00	92.40	90.54	95.95	92.30	90.01

V1 (Pose1): frontal view of face image, V2 and V3 (Pose2 and 3) multipose face images.