A team of scientists from Stanford University has developed an algorithm that allows you to modify facial expressions and lip movement on the already existing video material. And it does it so perfectly that it is almost impossible to distinguish a fake from the original film. This is another time when fantasy turns into reality.
Like us on Facebook you will always be up to date.
Artificial intelligence keeps getting better
The method developed by scientists from Stanford University uses the so-called deep learning. It is a type of machine learning using large-scale neural networks. What distinguishes it is the element of self-learning - no human intervention is required in this process. Deep learning has so far been used, for example, in speech recognition systems, machine translation or to recognize objects in photos. It is also increasingly used in the automotive industry - it is this phenomenon that enables creation autonomous cars. It also enables the creation of highly conscious robots.
How does it look like?
The method developed by scientists from Stanford University was presented in the video material. A footage was used with an actor whose facial expressions, head position, etc. are converted into a digital image. On this basis, the system verifies the video transmission. Similar solutions have already been implemented in the editing of portrait materials. However, the effects were not as spectacular as with the latest method.
See how it works
The development of artificial intelligence
So far, the most successful and well-known method of implementation artificial intelligence there is the robot Sophia. It was manufactured by Hanson Robotics from Hong Kong. As she claims, it was activated on April 19, 2015, and her face is modeled on the actress Audrey Hepburn. She is a kind of celebrity - apart from learning how to talk and work with people, she travels all over the world and gives interviews. It is also the world's first robot with granted citizenship. Sophia obtained a passport in 2017 and thus became an official citizen of Saudi Arabia.
Sophia will perform at the Impact'18 conference on June 13-14, 2018!
A double-edged weapon
However, it is worth being aware of both the benefits and possible threats associated with development artificial intelligence. This technology can be used, for example, to create high-quality dubbings in any language without the participation of a human voice-over. WITH second page – it is easy to imagine the effects of editing, for example, the speech of the head of state using the latest method from Stanford University. Therefore, every effort should be made to ensure that artificial intelligence is used extremely judiciously.