Baidu open sources AI to determine individuals with out face masks


With the rising scare of the lethal coronavirus, firms in China are pushing arduous to restrict its unfold. In a single such effort, the nation’s main search engine Baidu has open-sourced an AI mannequin to detect individuals not sporting face masks.

As coronavirus can unfold by means of shut contact with an contaminated individual by way of their coughs, sneezes, or respiratory droplets, China has made it necessary to put on face masks in a number of areas. Persons are instructed to wear masks in public places resembling eating places, purchasing malls, and public transport.

[Learn: How Chinese tech giants are assisting in the battle against coronavirus with AI]

It’s fairly arduous for authorities to catch individuals not sporting masks in massive crowds. That’s the place Baidu’s instrument can be helpful. The corporate mentioned it had skilled the mannequin on greater than 100,000 pictures to realize prediction accuracy of 96.5%.

As many factories and companies in China have resumed work after the Lunar New Yr vacation, it’s necessary for firms to determine individuals not sporting masks to restrict the unfold of the virus.

Whereas Baidu claims this can be a one-of-a-kind AI mannequin, it’s not that distinctive given how the corporate has developed plently of face detection models up to now.

In line with the World Well being Group (WHO), over 51,000 confirmed cases have been registered in China, with loss of life toll crossing 1,600.

Final week, the Chinese language authorities launched an app to assist individuals know if they have come in close contact with anyone affected by the virus.

Firms in China try to offer instruments and funding to gas analysis to search out the lethal virus’ treatment. Earlier this month, Baidu and e-commerce large Alibaba open-sourced AI models for college and medical facilities to enhance prediction occasions for RNA construction of the virus.

Printed February 17, 2020 — 05:39 UTC

Source link

Facebook Comments

This website uses cookies to improve your experience. We'll assume you're ok with this, but you can opt-out if you wish. Accept Read More