Saliency Optimization Based on Compactness and Background-Prior

Abstract

Saliency detection has drawn increasing attention in the communities of computer vision and image processing. Recently, foreground compactness and background prior have been developed to enhance saliency detection. In this paper, we propose an effective saliency optimization scheme taking account the foreground compactness and background prior. First, a foreground compactness-based saliency detection algorithm is introduced, which integrates the center contrast and the compactness-fused representation of the Gaussian Mixture Models (GMMs)-decomposed soft abstraction. Second, a foreground-based background seeds selection algorithm is proposed to obtain the enhanced background prior based saliency, which can well alleviate the influence of the on-boundary objects to the final saliency in conventional background prior based saliency algorithms. At last, the problem of compactness and background prior-based saliency integration is formulated as a multi-objective optimization problem to obtain the optimal saliency. Extensive experiments on ASD and MSRA10K database demonstrate that the proposed method outperforms the state-of-the -art saliency detection methods.