PROJECT TITLE :

Reflectance and Illumination Recovery in the Wild

ABSTRACT:

The appearance of an object in an image encodes invaluable information about that object and the surrounding scene. Inferring object reflectance and scene illumination from a picture would facilitate us decode this data: reflectance will reveal necessary properties concerning the materials composing an object; the illumination can tell us, as an example, whether the scene is indoors or outdoors. Recovering reflectance and illumination from a single image in the real world, but, may be a difficult task. Real scenes illuminate objects from each visible direction and real objects vary greatly in reflectance behavior. Yet, the image formation method introduces ambiguities, like color constancy, that build reversing the method ill-posed. To address this problem, we have a tendency to propose a Bayesian framework for joint reflectance and illumination inference in the $64000 world. We tend to develop a reflectance model and priors that exactly capture the space of real-world object reflectance and a flexible illumination model which will represent real-world illumination with priors that combat the deleterious effects of image formation. We analyze the performance of our approach on a set of artificial information and demonstrate results on real-world scenes. These contributions enable reliable reflectance and illumination inference in the important world.