A paper posted on arxiv a few days ago by Fan et al. claims to have a similar level of accuracy (97.3% on LFW): http://arxiv.org/pdf/1403.2802.pdf http://arxiv.

A paper posted on arxiv a few days ago by Fan et al. claims to have a similar level of accuracy (97.3% on LFW): http://arxiv.org/pdf/1403.2802.pdf http://arxiv.org/pdf/1403.2802.pdf Both methods use deep neural networks, but have a lot of differences, e.g., the Fan et al. paper doesn't use a 3D face model.