End-to-End Saliency Mapping via Probability Distribution Prediction

Abstract

Most saliency estimation methods aim to explicitly model low-level conspicuity cues such as edges or blobs and may additionally incorporate top-down cues using face or text detection. Data-driven methods for training saliency mod- els using eye-fixation data are increasingly popular, par- ticularly with the introduction of large-scale datasets and deep architectures. However, current methods in this lat- ter paradigm use loss functions designed for classification or regression tasks whereas saliency estimation is evalu- ated on topographical maps. In this work, we introduce a new saliency map model which formulates a map as a generalized Bernoulli distribution. We then train a deep ar- chitecture to predict such maps using novel loss functions which pair the softmax activation function with measures designed to compute distances between probability distri- butions. We show in extensive experiments the effective- ness of such loss functions over standard ones on four pub- lic benchmark datasets, and demonstrate improved perfor- mance over state-of-the-art saliency methods.