EF-Net: A novel enhancement and fusion network for RGB-D saliency detection

作者：

Highlights：

• We propose a novel enhancement-and-fusion framework for effective RGB-D saliency detection.

• We propose a novel depth enhancement model to improve the quality of depth maps and an effective layer-wise aggregation module to fuse the features extracted from RGB images and enhanced depth maps.

• Extensive experiments on five RGB-D benchmark datasets demonstrate that our method outperforms 12 state-of-the-art saliency detection methods by a large margin.

摘要

•We propose a novel enhancement-and-fusion framework for effective RGB-D saliency detection.•We propose a novel depth enhancement model to improve the quality of depth maps and an effective layer-wise aggregation module to fuse the features extracted from RGB images and enhanced depth maps.•Extensive experiments on five RGB-D benchmark datasets demonstrate that our method outperforms 12 state-of-the-art saliency detection methods by a large margin.

论文关键词：Salient object detection,RGB-D image,Depth enhancement,Feature fusion

论文评审过程：Received 16 May 2020, Revised 2 October 2020, Accepted 29 October 2020, Available online 4 November 2020, Version of Record 30 January 2021.

论文官网地址：https://doi.org/10.1016/j.patcog.2020.107740