HiDAnet: RGB-D Salient Object Detection via Hierarchical Depth Awareness

TL;DR

HiDAnet introduces a hierarchical depth-aware network for RGB-D saliency detection, effectively leveraging multi-scale and multi-modal features to improve salient object localization, especially in challenging scenarios.

Contribution

The paper proposes a novel Hierarchical Depth Awareness network that uses multi-granularity attention and a unified cross dual-attention module for enhanced multi-modal and multi-level feature fusion.

Findings

Outperforms state-of-the-art methods on benchmark datasets

Effectively distinguishes objects with similar appearance but different depths

Utilizes multi-scale loss for better hierarchical feature learning

Abstract

RGB-D saliency detection aims to fuse multi-modal cues to accurately localize salient regions. Existing works often adopt attention modules for feature modeling, with few methods explicitly leveraging fine-grained details to merge with semantic cues. Thus, despite the auxiliary depth information, it is still challenging for existing models to distinguish objects with similar appearances but at distinct camera distances. In this paper, from a new perspective, we propose a novel Hierarchical Depth Awareness network (HiDAnet) for RGB-D saliency detection. Our motivation comes from the observation that the multi-granularity properties of geometric priors correlate well with the neural network hierarchies. To realize multi-modal and multi-level fusion, we first use a granularity-based attention scheme to strengthen the discriminatory power of RGB and depth features separately. Then we introduce a unified cross dual-attention module for multi-modal and multi-level fusion in a coarse-to-fine manner. The encoded multi-modal features are gradually aggregated into a shared decoder. Further, we exploit a multi-scale loss to take full advantage of the hierarchical information. Extensive experiments on challenging benchmark datasets demonstrate that our HiDAnet performs favorably over the state-of-the-art methods by large margins.