Saliency refinement: Towards a uniformly highlighted salient object
作者：
Highlights：
•
• We propose a novel refinement framework of saliency maps by using a graph-based random sampling (GBRS) and saliency optimization.
• Our framework efficiently refines an initial saliency map by uniformly highlighting a salient object.
• Salient objects uniformly highlighted by our framework can be reliably utilized in various applications such as proto-object extraction and image retargeting.
摘要
•We propose a novel refinement framework of saliency maps by using a graph-based random sampling (GBRS) and saliency optimization.•Our framework efficiently refines an initial saliency map by uniformly highlighting a salient object.•Salient objects uniformly highlighted by our framework can be reliably utilized in various applications such as proto-object extraction and image retargeting.
论文关键词：Salient object detection,Saliency refinement,A uniformly highlighted salient object,Nonlocal L0 optimization
论文评审过程：Received 28 March 2017, Revised 6 December 2017, Accepted 7 December 2017, Available online 11 December 2017, Version of Record 20 December 2017.
论文官网地址：https://doi.org/10.1016/j.image.2017.12.003
