Saliency Detection Based on Background Priori and Target Priori
Author:Hui Kai, Lin Xiao, Liu Zu Xiang Author Unit: The College of Information, Mechanical and Electrical Engineering, Shanghai normal University, Shanghai 200234
Abstract:Image saliency detection is very useful in many computer vision tasks, but it is still a challenging problem.
This paper proposed a saliency detection method based on background prior and target prior. Firstly, we take
super-pixels outside the convex hull as background template to construct a background priori map. Then, a
Gaussian filtering model based on target paranoia is built in the convex hull. The model is combined with the
background priori map to obtain the target priori map. Finally, the background prior map and the target prior
map are fused and optimized by the Bayesian model. Experimental results on four public available datasets
show the effectiveness of the proposed method and the strength of our saliency map compared with several
previous method.