This neural network turns normal selfie in the 3D-model

The tool is developed by researchers at Nottingham and Kingston University, located in the UK. As employees told in his workLarge Pose 3D Face Reconstruction from a Single Image via Direct Volumetric CNN Regression.Usually to create a 3D-model of the head takes a few photos taken from different angles. However, feed the neural network a set of images and the corresponding three-dimensional models, the researchers were able to teach artificial intelligence to do this entire job using one photo.

The technology can be very useful. With it, for example, you could create 3D-avatars for video games and social networks in augmented reality. However, we can talk not only about individuals, but also any other objects.

As for the experimental instrument, as it can be slow - the author is clearly not enough server capacity. You may have to reload the page several times before you will be able to feed the neural network your picture.

Try the tool →