Multi-scale Cooperative Ranking for Saliency Detection
Saliency detection is an active research problem in computer vision and has been widely used in many applications.In this paper,we propose a new effective multi-scale cooperative ranking(MSR)model for image saliency detection.Our method begins with partitioning the input image into a set of super-pixels and constructing a neighborhood graph with super-pixels as nodes,both of which are performed at multiple scales of the input image,respectively.Then,we perform our MSR for super-pixels of different scales simultaneously and consistently with foreground cues or background cues as queries.MSR has a closed-form solution and thus can be computed efficiently.Experimental results on several benchmark databases show the effectiveness of the proposed MSR method.
Bo Jiang Xingyue Jiang Aihua Zheng Yun Xiao Jin Tang
School of Computer Science and Technology,Anhui University,No.111 Jiulong Road,Hefei,China
国际会议
广州
英文
574-585
2018-11-23（万方平台首次上网日期，不代表论文的发表时间）