Disney's New Eye-Animation Method Is Scary Good

Researchers at Disney have developed a new image-capture system that can reconstruct all the visible parts of a person's eye in animated form, at unprecedented levels detail – and the end result is damn impressive.


Eyes are notoriously difficult to render convincingly. Not only are they incredibly detailed, they're detailed in unique ways, varying dramatically from person to person. What's more, most of us, whether we realize it or not, are informal experts on the appearance of real, human eyes. Just think of how much time you spend looking at other people's eyeballs. Unless you're in the business of creating artificial eyes (and if you're not familiar with this process, I highly recommend checking it out), you probably can't name all the parts that make up an eye, or list the visual features that make them appear unique – but you can probably still tell a real one from a fake one, just based on the sheer number of eyes you've seen in your lifetime.

The upshot: If you want to animate an eye convincingly, you need to not only know what makes an eye look like an eye, but be able to reproduce all the visual qualities that correspond to those constituent parts. Or, as Markus Gross and his colleagues at Disney Research Zurich put it in their new publication:

Even though the human eye is one of the central features of individual appearance, its shape has so far been mostly approximated in our community with gross simplifications. In this paper we demonstrate that there is a lot of individuality to every eye, a fact that common practices for 3D eye generation do not consider. To faithfully reproduce all the intricacies of the human eye we propose a novel capture system that is capable of accurately reconstructing all the visible parts of the eye: the white sclera, the transparent cornea and the non-rigidly deforming colored iris. These components exhibit very different appearance properties and thus we propose a hybrid reconstruction method that addresses them individually while respecting their inter-dependencies. The result of our system is a complete model of both spatio-temporal shape and texture at an unprecedented level of detail. The captured eyes will not only enable the creation of more believable digital humans but have potential applications in fields other than computer graphics, such as ophthalmology. Finally, we believe that the findings of this paper will alter our community's current assumptions regarding human eyes, and our work has the potential to significantly impact the way that eyes will be modelled in the future.

Damn. Those are some beautifully animated peepers.

Read the full writeup over at Disney Research.



Here's a compilation of iris photographs by Suren Manvelyan.

Irises freak me out.