The prototype for the virtual host was a real person.
Chinese news Agency Xinhua has presented leading news, which was completely modeled by a system of artificial intelligence.
Virtual presenters got the image of a real speaker Zhang Zhao and read the news using synthesized voices.
Engineers have created two leading: one for the English broadcast, and one for Chinese. In Xinhua saying that such images have “infinite possibilities”, and they are much cheaper than the leading of flesh and blood.
It is unclear what technology was used to create the images. Presumably, the engineers from Xinhua news used footage of Zhang Zhao as a base layer, and then animated the parts of the mouth and face, to turn the virtual presenter. By combining this solution with the synthesized voice, Xinhua news may obtain e-news anchor.
“Each presenter can work 24 hours a day on the official website and on various social networking platforms, reducing the costs of news production and improving efficiency” – assured Xinhua.
However, technology has its limitations. The rollers can be seen that the range of facial expressions new leading limited, and voice is clearly artificial. But research in the field of machine learning in this area is rapidly improving, and it is easy to imagine a future where the virtual leading is indistinguishable from the real one.
“The development of the media industry requires constant innovation and deep integration of advanced technologies. I will work tirelessly to keep you up to date, as the texts will come into my system,” he told the virtual host to its first broadcast.
© 2018, z-news.link. All rights reserved.