Saliency propagation with perceptual cues and background-excluded seeds

Graph-based methods have shown their potentialities for saliency detection. In this paper, a graph-based framework is proposed for saliency detection, which incorporates perceptual cues into the framework and uses the background-excluded seeds to propagate saliency. Firstly, a graph is constructed by two perceptual cues, including proximity and similarity. Secondly, probable background nodes are generated by a novel background probability measure and used to pick out reliable seeds. Then a label propagation model is developed to diffuse saliency based on these reliable seeds. Lastly, another perceptual cue called rareness is integrated into a cost function to optimize the propagation result. Results on four datasets demonstrate that the proposed method achieves superior performance against fifteen state-of-the-art methods in terms of different evaluation metrics.