SALIENT REGION DETECTION USING BACKGROUND AND FOREGROUND PRIORS
张云 · 2013
In this paper, we present a new saliency detection method which incorporates foreground and background priors. We consider the center region as the foreground prior and the border as the background prior. A unique saliency measurement is given which measures the dissimilarity from the background region versus the similarity to the foreground region. The final saliency map is obtained through iterative refinement starting from the priors. Experimental results demonstrated that the proposed algorithm, while simple, is robust in terms of the salient object location and outperforms many state-of-the-art methods on the public benchmark dataset.