Yin Shiming, vice-president of Baidu, presides over a cloud computing sub-forum at the Baidu Create 2017 conference on July 5, 2017. (Photo by Song Jingli/chinadaily.com.cn)
Baidu's artificial intelligence not only makes driving hands-free possible, but could also let machines analyze video footages much better than humans, the company's executives shared on a developers' conference in Beijing Wednesday.
"It might be impossible for you to figure out whether your son said 'Daddy, I love you' when he was two-year-old in videos, but it is possible now," Yin Shiming, vice-president of Baidu, told audience at a cloud computing sub-forum of the Baidu Create 2017 conference.
Huang Jingbo, a research manager at Baidu Cloud's multimedia department, said that Baidu has rolled out video processing services such as video content analysis (VCA) and smart selection of video covers.
When the VCA service, enabled by deep learning, is applied on a footage, it will be tagged automatically and elements of the footage, including voice, characters, scenarios, faces will all be analyzed and displayed, Huang told the audience.
He added that the smart cover selection function, partly enabled by machines' understanding of human feelings, could recommend a video to its possible viewers in the best way, rather than always the first frame of the video.
Yin, also the general manger of Baidu Cloud, said Baidu's advancement in visual sector will impact many sectors, with entertainment and education as the most possible.