It looks like 3D Selfies will come very soon thanks to a technology developed by a group of scientists from the universities of Nottingham and Kingston.

Unlike how it is currently done 3D reconstruction of the image of a face where you normally need photos from various angles of the same and the machine must establish parallels and concordances between images through complex algorithms, the team of scientists decided to use another method. They decided to train a Neural Convolutional Network (CNN) with 2D images of faces and scanning or 3D facial models of them.

Thanks to being convolucional has memory therefore this CNN ...

... works with just a single 2D facial image, does not require accurate alignment nor establishes dense correspondence between images, works for arbitrary facial poses and expressions, and can be used to reconstruct the whole 3D facial geometry (including the non-visible parts of the face) bypassing the construction (during training) and fitting (during testing) of a 3D Morphable Model. ... performs direct regression of a volumetric representation of the 3D facial geometry from a single 2D image.

They have a page where you can upload your own image and create your face in 3D as I did with the photo of Obama at the beginning of the publication (that photo is given as an example on the page). From what I have been able to see the modeling is super fast and also give you the option to download the OBJ file from the 3D construction of the face.

I am sure that very soon we will have the option of 3D Selfies on the phones. I see it coming.

To learn more about this new technology of 3D scanning from a 2D photo here I leave the link to the website.