The popularity of social networks has brought the rapid growth of social images which have become an increasingly important\nimage type. One of the most obvious attributes of social images is the tag. However, the sate-of-the-art methods fail to fully\nexploit the tag information for saliency detection.Thus this paper focuses on salient region detection of social images using both\nimage appearance features and image tag cues. First, a deep convolution neural network is built, which considers both appearance\nfeatures and tag features. Second, tag neighbor and appearance neighbor based saliency aggregation terms are added to the saliency\nmodel to enhance salient regions.The aggregation method is dependent on individual images and considers the performance gaps\nappropriately. Finally, we also have constructed a new large dataset of challenging social images and pixel-wise saliency annotations\nto promote further researches and evaluations of visual saliency models. Extensive experiments show that the proposed method\nperforms well on not only the new dataset but also several state-of-the-art saliency datasets.
Loading....