Samsung AI Center Shows New Algorithm That Can Animate a Talking Head With Just a Single Photo

Artificial intelligence or AI can be a useful tool to help take the burden off our shoulders when it comes to trivial and mundane tasks.

Companies like Google have employed AI and machine learning to emulate individuals for simple tasks such as reserving a salon appointment, with products like google Duplex. AI calculations have also been used to animate 2D objects such as talking heads of people, but this typically requires a huge dataset of samples to train the neural network to accommodate to the objects or in this instance, that the individual’s behavior and mannerisms. This process just got a whole lot simpler thanks to Samsung AI Center’s Moscow chapter, which has published a paper that shows that a highly realistic talking head model can be obtained by feeding the algorithm together with as little as one sample photo.

The algorithm utilizes a method of utilizing”landmarks,” that is a skeletal frame of a persons face learnt by crawling through a large dataset of movies, which is used as the backbone to its synthesised face. The precision of the synthesised confront depends on the number of source images you feed the algorithm.

The newspaper indicates the gap between models trained using 32 samples, eight samples, and one sample. The output from just 1 sample is obviously has a couple of artefacts around the face, helping you spot that it is not a genuine video but an AI-generated one. However, the fact we can even get such output with one sample remains shocking. The video and the paper indicates the algorithm successfully producing synthesised faces of actors and even paintings, like the Mona Lisa.

Such a technique can be used to deliver you more realistic looking’Memojis,’ that can be Apple’s version of an Animoji but of yourself. The creative potential of the is endless. However, such technology may also be misused if fallen in the wrong hands. Stars have often become the target of these”deepfake,” videos and images, which is aimed to tarnish their social reputation. It may of course, also be used to make fake news.

Before this season, Hollywood actress Scarlett Johansson was vocal against the dangers these”deepfakes,” after her face was grafted into pornographic videos. Making such fake photos and videos isn’t a new dilemma, but the fact that it may be achieved this easily now, is very worrying. Social networking platforms such as Twitter, Reddit and even adult entertainment portals such as Pornhub have begun cracking down of AI-generated”deepfakes.”