How many bits does it take for a stimulus to be salient?

Visual saliency has been shown to depend on the unpre-dictability of the visual stimulus given its surround. Var-ious previous works have advocated the equivalence be-tween stimulus saliency and uncompressibility. We pro-pose a direct measure of this quantity, namely the number of bits required by an optimal video compressor to encode a given video patch, and show that features derived from this measure are highly predictive of eye fixations. To ac-count for global saliency effects, these are embedded in a Markov random field model. The resulting saliency mea-sure is shown to achieve state-of-the-art accuracy for the prediction of fixations, at a very low computational cost. Since most modern cameras incorporate video encoders, this paves the way for in-camera saliency estimation, which could be useful in a variety of computer vision applications. 1