ZhigangTU/Spatiotemporal-Salient-Object-Detection-in-Video

We present a spatiotemporal saliency detection method to extract salient objects. Unlike previous methods focusing on exploiting or incorporating different features, the proposed method integrates two distinctive saliency maps, which are respectively computed from an appearance-dominated method and a motion-dominated algorithm captured object candidates, to obtain a refined spatiotemporal saliency maps.