Attentive Cross-Modal Fusion Network for RGB-D Saliency Detection

In this paper, an attentive cross-modal fusion (ACMF) network is proposed for RGB-D salient object detection. The proposed method selectively fuses features in a cross-modal manner and uses a fusion refinement module to fuse output features from different resolutions. Our attentive cross-modal fusion network is built based on the residual attention. In each level of ResNet output, both the RGB and depth features are turned into an identity map and a weighted attention map. The identity map is reweighted by the attention map of the paired modal. Moreover, the lower level features with higher resolution are adopted to refine the boundary of detected targets. The whole implementation can be trained end-to-end. Our experimental results show that the ACMF exceeds state-of-the-art methods on five recent datasets for RGB-D salient object detection with averagely 9.0% gain in F-measure, 6.7% gain in S-measure, and 37.2% reduction in MAE.