The Taub Faculty of Computer Science Events and Talks

Photographs of hazy scenes typically have low-contrast and offer a limited scene visibility. I will describe a new method for single-image dehazing that relies on a generic regularity in natural images in which pixels of small image patches exhibit one-dimensional distributions in pixel space. I will derive a local formation model that explains this formation in the context of hazy scenes and use it for recovering the scene transmission based on the models' offset. Moreover, this model allows identifying and dismissing pixels that do not follow it and hence, unlike existing approaches that follow their assumptions across the entire image, the new algorithm validates its hypotheses and obtains more reliable estimates where possible. I will also describe a Markov random field model that is dedicated for producing complete and regularized transmission maps. Unlike traditional field models that consist of local coupling, the new model is augmented with long-range connections between pixels of similar color. This allows the algorithm to properly resolve the transmission in isolated regions where nearby pixels do not offer relevant information. An extensive evaluation of the new method over different types of images and its comparison to state-of-the-art methods on established benchmark images show a consistent improvement in the accuracy of the estimated scene transmission and recovered haze-free radiances.