3D data generation

While it's easy for a human to deduce the 3D shape of an object from a single perspective, it's extremely difficult for an algorithm. Indeed, unlike computer vision, human vision is intensely aided by memory and intelligence.

Yet there is a demand for the automatic extraction of high-quality 3D objects from a single photograph. The applications are numerous, ranging from image manipulation to virtual and augmented reality, not to mention 3D printing. For example, in e-commerce, it would be very interesting to be able to automatically and rapidly retrieve the 3D model of a commercial product from its photograph (for example, in advertising). In addition, geometry and texture must be of high quality to be useful. In what follows, we will present work that has enabled qualitative advances towards this goal.

Limited by the nature of the deep...