Saliency Research Via Random Walk And Angular Embedding

In recent years,salient object detection has been applied to many computer vision tasks,such as image retrieval,image segmentation,image classification,object recognition,to name a few.Although much significant progress has been made,salient object detection remains a challenging problem.In this paper,we propose a bottom-up saliency model based on random walk(RW).Firstly,a sparsely connected graph is constructed to capture the local context information of each node.All image boundary nodes and other nodes are respectively treated as the absorbing nodes and transient nodes in the absorbing Markov chain.Then the expected number of times from each transient node to all other transient nodes can be used to represent the saliency value of this node.The absorbed time depends on the weights on the path and their spatial coordinates,which are completely encoded in the transition probability matrix.Considering the importance of this matrix,we adopt different hierarchies of deep features extracted from fully convolutional networks(FCN)and learn a transition probability matrix,which is called learnt transition probability matrix.Although the performance is significantly promoted,salient objects are not uniformly highlighted very well.To solve this problem,angular embedding(AE)technique,whose goal is to find the global orderings of all elements where their relative differences between elements match well the pairwise local ordering measurements,is investigated to refine the saliency results.Based on pairwise local orderings,which are produced by the saliency maps of RW and boundary maps,we rearrange the global orderings(saliency value)of all nodes.Extensive experiments demonstrate that the proposed algorithm outperforms the state-of-the-art methods on six publicly available benchmark datasets.