Salient video object detection using a virtual border and guided filter

作者：

Highlights：

• Detect the salient object connected to frame borders using the distance transform with a virtual border-based technique.

• Introduce an edge-aware filter to fuse the spatial edge with the temporal information to enhance salient object edges.

• Compute Entropy and Standard deviations to decide the confidence level of the spatial saliency map and the temporal saliency map.

摘要

•Detect the salient object connected to frame borders using the distance transform with a virtual border-based technique.•Introduce an edge-aware filter to fuse the spatial edge with the temporal information to enhance salient object edges.•Compute Entropy and Standard deviations to decide the confidence level of the spatial saliency map and the temporal saliency map.

论文关键词：Video salient object detection,Distance transform,Guided filter,Global motion

论文评审过程：Received 21 December 2017, Revised 31 May 2019, Accepted 7 August 2019, Available online 7 August 2019, Version of Record 19 August 2019.

论文官网地址：https://doi.org/10.1016/j.patcog.2019.106998