Skip to main content

Table 4 Verification accuracy (%) comparison on the CFP dataset

From: Facial UV map completion for pose-invariant face recognition: a novel adversarial approach based on coupled attention residual UNets

Model

Frontal–frontal

Frontal-profile

CASIA-center [35]

\(98.34 \pm 0.44\)

\(87.77 \pm 2.39\)

CASIA-Sphere [35]

\(98.64 \pm 0.24\)

\(84.39 \pm 2.59\)

CASIA-sm [35]

\(98.59 \pm 0.21\)

\(87.74 \pm 1.07\)

CASIA-sm-aug1 [35]

\(98.25 \pm 0.42\)

\(90.14 \pm 1.53\)

CASIA-sm-augUV-GAN [35]

\(98.83 \pm 0.27\)

\(93.09 \pm 1.72\)

 Profile2Frontal

–

\(93.55 \pm 1.67\)

 Frontal2Profile

–

\(93.72 \pm 1.59\)

 Template2Template

–

\(94.05 \pm 1.73\)

CASIA-arcface-augAttentionResCUNetGAN

\(99.47 \pm 0.29\)

\(97.0 \pm 0.82\)