Source: Global Network
[Global Network Technology Comprehensive Report] On November 20, Digital Kingdom Group Co., Ltd. (hereinafter referred to as "Digital Kingdom"), a well-known visual special effects and virtual reality experience service provider, showed its advanced virtual person - "Douglas" for the first time in the recently launched "Intel's AI Podcast".
is different from the virtual assistants and chat robots that use pure voice as the interaction mode. Digital Kingdom chooses to use realistic digital stand-alone, namely the virtual human "Douglas", which is the carrier of human-computer interaction. Digital Kingdom said that Douglas has facial features that are no different from real humans and has the ability to change appearance like a chameleon, which gives it a wide and flexible possibility to enter the consumer-level path in 2021.
It is understood that Douglas evolved from a digital stand-alone created by Dr. Doug Roble, senior director of the Digital Kingdom software research and development department. Previously, related technologies have made many achievements in building virtual people that can be manipulated in real time and facial capture integrated into artificial intelligence.
To create Douglas, Doug Roble performed hundreds of hours of performance capture, including reading books aloud to record expressions and sounds. Digital Kingdom then uses portrait photos taken in a variety of lighting conditions to train neural rendering tools. Relying on captured data, the tool can now achieve a level of reality that traditional technologies cannot achieve, while making the process of generating digital stand-alones faster. It is said that even with only a few expressions and sounds, such as videos with a duration of about 10 minutes or audios with about 30 minutes, can accurately imitate another person's behavior.
According to reports, Douglas is based on machine learning solutions and independently developed technology by Digital Kingdom, which is enough to simulate the actions of real humans. Through language processing, expression sculpting and visual tracking, it can also recognize and talk to humans. It is said that its dialogue level is comparable to that of smart voice assistants "Alexa" and "Siri", and will not have long pauses during communication, and can communicate smoothly and naturally with humans.
Therefore, in addition to facial realistic, Douglas is also particularly suitable for application scenarios composed of repetitive tasks, such as customer service areas that require multiple times to deal with the same problem. The current version of Douglas can be connected to any virtual assistant or chat robot system , providing virtual face and emotional intelligent responses for real-time human-computer interaction. The refined Douglas is expected to support the deployment of online conference platforms worldwide. As a virtual person, its realistic effect indicates that this technology may leverage more diverse sensory experiences.
Digital Kingdom CEO Xie An said: "The birth of Douglas is of great significance, which marks an unprecedented autonomous virtual person that is about to embrace the consumer-grade human-computer interaction track with huge capacity and urgent need to be released. Artists and scientists from Digital Kingdom continue to devote themselves to the expansion of facial capture and autonomous behavioral manipulation, and accelerate the commercialization process of virtual people to the consumer group."
