Multiple feature distinctions based saliency flow model

作者：

Highlights：

• A new saliency flow model is proposed to formulate the saliency spread process.

• Multiple feature distinctions are used to overcome the disturbance of background.

• Both low level and mid level features are considered.

• The proposed method represents higher robustness in the complex scenes.

摘要

Highlights•A new saliency flow model is proposed to formulate the saliency spread process.•Multiple feature distinctions are used to overcome the disturbance of background.•Both low level and mid level features are considered.•The proposed method represents higher robustness in the complex scenes.

论文关键词：Salient object detection,Multiple features distinction,Saliency flow,Quadratic programming

论文评审过程：Received 24 August 2015, Revised 26 November 2015, Accepted 19 December 2015, Available online 31 December 2015, Version of Record 27 February 2016.

论文官网地址：https://doi.org/10.1016/j.patcog.2015.12.014