By super nervous
Sa Beining, one of the hosts sharing the stage with AI surrogates, announced on the spot that he was ready to “retire”. Is the moment when AI will replace the host really coming?
In this year’s network spring evening, the first appearance of four virtual image hosts, respectively based on sa Beining, Zhu Xun, Gao Bo, Long Yang to create. The four hosts all share the stage with their own virtual hosts, which really adds a lot of fun to the evening.
Jump to watch: four virtual hosts
It can be seen from the video that there is a great similarity between the virtual AI host and the entity. In addition to answering and shaking, the AI host has corresponding body language, and each virtual host also has different expression characteristics.
What is a virtual host
It is reported that the virtual host of this online Spring Festival Gala is PAI (Personal AI) created by ObEN, which can create a virtual mirror model based ona person’s image.
The core technology of this product is mainly in three aspects:
First, the establishment of visual image. Using the photos of the host body, 3d scanner, 3DS MAX, MAYA and other 3D modeling tools, let the AI build an algorithm model, by understanding the relationship between the color distribution and the structural depth of the photos, finally build the face and body shape.
Second, there is sound synthesis. The AI speech technology they use does not need to collect a large number of speech libraries, but only need to record more than ten sentences of speech. Through feature parameter extraction, transfer learning and other methods, the sound model can be established.
Finally, make the virtual host as realistic as possible. This requires not only the visuals to match their sounds, but also the personalization of them. Through sensors and motion tracking devices, combined with AI and motion capture training, virtual hosts can simulate oral expressions, facial expressions, gestures, body movements and scene interaction according to the corresponding ontology.
Although this is the first time AI host crossover, but AI virtual anchor industry is really not a few appearances.
The first AI news anchor
At the fifth Internet Conference held in November 2018, Xinhua News Agency unveiled a new reporter, who was the first news anchor synthesized by ARTIFICIAL intelligence (AI). Qiu Hao, the prototype of AI, said, “The image is my image, and the voice is also suitable for my voice, but I have never said…….”
Jump to watch, AI virtual anchor
In this video, we can see that the host in the video introduces himself with his voice, and his face and lips have corresponding movements in accordance with the voice. I have to say, surprise, but disappointment. It can be easily identified as a product of mechanization rather than a real person. It works in a way that requires humans to write press releases. but
This kind of technical support, come from sogou “double body technology”. The key to this technology is speech synthesis and image generation.
Speech synthesis technology, through a small amount of audio data, the machine model can learn the speech characteristics of the input object, grasp the timbre, rhythm, emotion and other aspects, and finally realize the audio information of the input text.
And image generation, using face recognition, three-dimensional face reconstruction, expression modeling and other aspects of learning and construction, and finally achieve the output of audio output visual information corresponding.
Despite the drawbacks, one of the biggest claims of the technology is its on-screen versatility, which is basically the ability to watch one person broadcast three different things at once on your TV, tablet or phone.
Earlier AI anchors
And Japan had reported earlier, they also made AI anchor.
In April last year, Japan’s NHK aired a program called “NEWS CHECK 11” featuring a cute cartoon AI anchor named “NEWS Anchor Yomiko.”
This anchor is made by using CG technology. It learns to divide phonemes from a large number of recordings, and then learns to recognize and read texts. Finally, it is used to read news.
In addition, robots as news anchors have been realized.
Elika, a beauty robot developed by Osaka University and Kyoto University, also became a news anchor of NNN in Japan in April 2018.
Elika, who is 23 years old, is endowed with a typical female face and speaks in a natural voice based on recordings of voice actors.
In addition, it also has an advanced dialogue system. When talking with a person, it collects information through microphones and sensors, perceives the other person’s voice and movements, and turns to the other person for a smooth conversation.
Its eyes, mouth, neck and other 19 places can move by air pressure, presenting a variety of expressions, but also some simple movements, lifelike.
When the host will be replaced
Back to the Online Spring Festival Gala, the appearances of the four mischievous virtual hosts received a lot of attention, and the audience’s attitude showed that they liked them.
So could they be an opportunity to lay off presenters? Well, maybe neither Saz nor Saz would say yes.
As for the tone and disobedient facial expressions of AI news live, there is still room for improvement. Japan’s once-popular news robots have not largely replaced local journalism, at best supporting and novel.
So in that sense, technological advances do bring us novelty and beautiful experiences, but perhaps in an age when gimmicks move faster than reality.
Maybe the day will come, but it certainly won’t be today. We should believe that by that time, humans may have already figured out how to get along with AI. At that time, we will see AI not only hosting the Spring Festival Gala, but also performing programs. We ah, open mouth waiting for AI to feed us.
Maybe the day will come, but it certainly won’t be today. We should believe that by that time, humans may have already figured out how to get along with AI. At that time, we will see AI not only hosting the Spring Festival Gala, but also performing programs.
We ah, open mouth waiting for AI to feed us.