Abstract:

The digital photography revolution has greatly facilitated the way in
which we take and share pictures. However, it has mostly relied on a
rigid imaging model inherited from traditional photography.
Computational photography goes one step further and exploits digital
technology to enable arbitrary computation between the light array and
the final image or video. Such computation can overcome limitations of
the imaging hardware and enable new applications. It can also enable new
imaging setups and post-processing tools that empower users to enhance
and interact with their images and videos.

This talk will describe new imaging architectures as well as software
techniques that leverage computation to facilitate the extraction of
information and enhance images. In particular, I will discuss post
exposure manipulations like reflections removal, colorization and
matting. I will also describe the coded aperture camera, a new simple
modification of a lens as well as new inference techniques that enable
the capture of both depth and a full-resolution image from a single
picture.
I will argue that the core of computational photography research is in
the fact that images are more than arbitrary random arrays of numbers,
and the success of such algorithms depends on the ability to model the
strong low level statistics of images.


Parts of this research were done in collaboration with Fredo Durand, Rob
Fergus, Bill Freeman, Dani Lischinski, Alex Rav Acha, Yair Weiss and
Assaf Zomet