A prior regularized multi-layer graph ranking model for image saliency computation

Bottom-up saliency detection has been widely studied in many applications, such as image retrieval, object recognition, image compression and so on. Saliency detection via manifold ranking (MR) can identify the most salient and important area from an image efficiently. One limitation of the MR model is that it fails to consider the prior information in its ranking process. To overcome this limitation, we propose a prior regularized multi-layer graph ranking model (RegMR), which uses the prior calculating by boundary connectivity. We employ the foreground possibility in the first stage and background possibility in the second stage based on a multi-layer graph. We compare our model with fifteen state-of-the-art methods. Experiments show that our model performs well than all other methods on four public databases on PR-curves, F-measure and so on.