Unsupervised image saliency detection with Gestalt-laws guided optimization and visual attention based refinement

作者：

Highlights：

• Gestalt laws guided saliency detection via characterizing HVS and forming objects.

• Smooth at superpixel and object levels by fusing bottom-up and top-down mechanisms;

• Background suppression with background correlation term & spatial compactness term.

• Two-stage refinement to show best among 10 state-of-the-art methods on 5 datasets.

摘要

•Gestalt laws guided saliency detection via characterizing HVS and forming objects.•Smooth at superpixel and object levels by fusing bottom-up and top-down mechanisms;•Background suppression with background correlation term & spatial compactness term.•Two-stage refinement to show best among 10 state-of-the-art methods on 5 datasets.

论文关键词：Background connectivity,Gestalt laws guided optimization,Image saliency detection,Feature fusion,Human vision perception

论文评审过程：Received 29 May 2017, Revised 16 January 2018, Accepted 2 February 2018, Available online 5 February 2018, Version of Record 14 February 2018.

论文官网地址：https://doi.org/10.1016/j.patcog.2018.02.004