基于图的流形排序的多层级融合显著性目标检测

In this paper, we form the saliency map by a multi-level combination approach using graph-based manifold ranking method as the foundation. This method aims to overcome the defects that the merely graph-based manifold ranking method can not do well in highlighting the salient objects uniformly and suppressing the background effectively. We execute over-segmentation algorithm for each input image, and in order to make full use of the image information and cues, we segment the input image into multilevels from the coarsest to the finest according to the number of segmented regions. And for each level, graph-based manifold ranking algorithm is used to generate the initial saliency map. Then we obtain the optimal weight of each level from the trained combiner to make a multi-level combination to get the final saliency map. The combiner is trained using quadratic programming algorithm on the MSRA-10K database. In the experiment, we compare our method with the other eight state-of-the-art methods in terms of four evaluation criteria. The experimental results demonstrate that the proposed method achieves consistent and favorable performance against the eight state-of-the-art methods on different databases.?????