No abstract

Computational saliency refers to the ability to highlight the salient visual information for processing. The mechanism has proven to be helpful for human as well as computer vision. Computational saliency focuses on designing algorithms which, similarly to human vision, predict which regions in a scene are salient. Recently, salient object segmentation has introduced the use of object proposals. Object proposal methods provide image segments as proposals which can be used for saliency estimation. We propose several saliency features which are computed from different networks and different levels with the aim to define which optimal network and layer for the task of saliency detection. Also much recently, convolutional neural networks breakthroughs computer vision with the extraction of the much powerful features which are based on deep CNN. In this paper, we develop a saliency approach based on the computation of deep whitened features combined with shape features from object proposals. We train an SVM to predict the saliency of every object proposal. Experimental results shows that we outperform other state-of-the-art methods in PASCAL-S, FT, ECSSD, MSRA-B and ImgSal data sets in terms of F-score, PR curves. Furthermore, experiments show that applying whitening improve performance.

No abstract

No abstract

Top-down semantic information is known to play an important role in assigning saliency. Recently, large strides have been made in improving state-of-the-art semantic image understanding in the fields of object detection and semantic segmentation. Therefore, since these methods have now reached a high-level of maturity, evaluation of the impact of high-level image understanding on saliency estimation is now feasible. We propose several saliency features which are computed from object detection and semantic segmentation results. We combine these features with a standard baseline method for saliency detection to evaluate their importance. Experiments demonstrate that the proposed features derived from object detection and semantic segmentation improve saliency estimation significantly. Moreover, they show that our method obtains state-of-theart results on (FT, ImgSal, and SOD datasets) and obtains competitive results on four other datasets (ECSSD, PASCAL-S, MSRA-B, and HKU-IS).