Progressive multi-scale fusion network for RGB-D salient object detection

作者：

Highlights：

• Novel multi-scale structure for RGB-D saliency detection.

• Mask-Guided Feature Aggregation module for filtering noise in depth data.

• Mask-Guided Refinement Module for filtering noise from multi-scale RGB data.

• Progressive fusion strategy from deep to shallow layers.

• Achieve competitive performance compared to 11 prevalent methods.

摘要

•Novel multi-scale structure for RGB-D saliency detection.•Mask-Guided Feature Aggregation module for filtering noise in depth data.•Mask-Guided Refinement Module for filtering noise from multi-scale RGB data.•Progressive fusion strategy from deep to shallow layers.•Achieve competitive performance compared to 11 prevalent methods.

论文关键词：Multi-scale fusion,Mask guided,Salient object detection

论文评审过程：Received 27 January 2022, Revised 21 July 2022, Accepted 3 August 2022, Available online 15 August 2022, Version of Record 24 August 2022.

论文官网地址：https://doi.org/10.1016/j.cviu.2022.103529