Got two left feet? New AI can transfer a professional dancer’s moves onto YOUR body

A new AI can make anybody look like a dance superstar by giving them the moves of a professional dancer.

The creepy robot uses the movements of top dancers captured from a video to animate your body in a creepy ‘DeepFake’-style video.

Experts behind the technology said it allows amateurs to ‘twirl like ballerinas’, or perfectly imitate their favourite pop stars.

 

A new AI can make anybody look like a dance superstar by giving them the moves of a professional dancer. The creepy robot uses the movements of top dancers (left) captured from a video to animate your body (right images) in a creepy ‘DeepFake’-style video

The AI, developed by researchers at the University of California at Berkeley, first copies the moves of a professional dancer from a source video.

It then generates a new video of a completely different person’s body performing those moves.

The films it generates feature occasional blurs and the odd missing body part, but reveal the potential for the technology to one day produce convincing video.

‘With our framework, we create a variety of videos, enabling untrained amateurs to spin and twirl like ballerinas, perform martial arts kicks or dance as vibrantly as pop stars,’ researchers wrote in their paper.

The system learns to transfer dance moves using deep learning, a type of machine intelligence that pits several AIs against one another.

One AI creates a virtual skeleton on which to map poses, while two more challenge one another to create more realistic video.

One AI, a generator, produces a video, while a discriminator tries to determine if the image is real or fake, sending the clip back to the first AI if it isn’t realistic enough.

Over thousands of calculations the two AIs eventually produce a realistic video in which the moves of one subject are mapped onto the frame of another.

The team had to capture a short video of the non-dancer twisting around to give the AI sample footage with which to build the fake clip. 

‘Given a video of a source person and another of a target person, our goal is to generate a new video of the target person enacting the same motions as the source,’ researchers wrote. 

The research is the latest step forward in creating realistic video that places people in situations that they were never really in.

Experts behind the technology said it allows amateurs to 'twirl like ballerinas', or perfectly imitate their favourite pop stars. Pictured are two fake videos (centre and right) thar made amateurs appear as if they were dancing like Bruno Mars (top left)

Experts behind the technology said it allows amateurs to ‘twirl like ballerinas’, or perfectly imitate their favourite pop stars. Pictured are two fake videos (centre and right) thar made amateurs appear as if they were dancing like Bruno Mars (top left)

HOW DOES ARTIFICIAL INTELLIGENCE LEARN?

AI systems rely on artificial neural networks (ANNs), which try to simulate the way the brain works in order to learn.

ANNs can be trained to recognise patterns in information – including speech, text data, or visual images – and are the basis for a large number of the developments in AI over recent years.

Conventional AI uses input to ‘teach’ an algorithm about a particular subject by feeding it massive amounts of information.   

AI systems rely on artificial neural networks (ANNs), which try to simulate the way the brain works in order to learn. ANNs can be trained to recognise patterns in information - including speech, text data, or visual images

AI systems rely on artificial neural networks (ANNs), which try to simulate the way the brain works in order to learn. ANNs can be trained to recognise patterns in information – including speech, text data, or visual images

Practical applications include Google’s language translation services, Facebook’s facial recognition software and Snapchat’s image altering live filters.

The process of inputting this data can be extremely time consuming, and is limited to one type of knowledge. 

A new breed of ANNs called Adversarial Neural Networks pits the wits of two AI bots against each other, which allows them to learn from each other. 

This approach is designed to speed up the process of learning, as well as refining the output created by AI systems. 

So-called DeepFake technology has already been used to create fabricated clips that transfers people’s faces and mouths to a video of someone saying or doing things they never did.

Last year internet users began utilising the technology to create fake videos in which the faces of celebrities were pasted onto porn stars.

PornHub, Reddit and Twitter all banned AI-generated porn from being posted on their platforms in February.

They said the content falls under the guidelines of non-consensual porn, which violates site-wide policies.

Read more at DailyMail.co.uk