Text: | Print|

Facial recognition software more accurate than human eye

2014-06-23 16:21 Ecns.cn Web Editor: Gu Liping

(ECNS) -- Researchers at the Chinese University of Hong Kong say they have invented software that can identify people from photographs better than humans can.

The team led by Professor Tang Xiaoou tested the system on a database known as Labelled Faces in the Wild (LFW), which has 13,000 photos of 6,000 public figures. In the test, humans and computers were asked to compare photos and determine whether they were the same person.

Humans performed the face-verification task with 97.53 percent accuracy, while the software DeepID achieved 99.15 percent.

The team had already created a system called GaussianFace, which achieved a success rate of 97.45 percent.

The new system ranks higher than Facebook's DeepFace technology.

Facial recognition technology has applications in security, law enforcement, mobile internet and entertainment.

Comments (0)
Most popular in 24h
  Archived Content
Media partners:

Copyright ©1999-2018 Chinanews.com. All rights reserved.
Reproduction in whole or in part without permission is prohibited.