Bio-inspired visual attention model and saliency guided object segmentation

Abstract

In this paper, we present a saliency guided image object segment method. We suppose that saliency maps can indicate informative regions, and filter out background in images. To produce perceptual satisfactory salient objects, we use our bio-inspired saliency measure which integrating three factors: dissimilarity, spatial distance and central bias to compute saliency map. Then the saliency map is used as the importance map in the salient object segment method. Experimental results demonstrate that our method outperforms previous saliency detection method, yielding higher precision (0.7669) and better recall rates (0.825), F-Measure (0.7545), when evaluated using one of the largest publicly available data sets.