Yuanlin Chen ^* Zengbao Sun Cheng Yan ^* Ming Zhao ^*

• Shanghai Maritime University, pudong, Shanghai, China

RGB-T Salient object detection (SOD) is to accurately segment salient regions in both visible light images and thermal infrared images. However, most of existing SOD methods neglects the critical complementarity between different modalities images, which is beneficial to further improve the detection accuracy. In this letter, we propose the edge-guided feature fusion network (EGFF-Net), which consists of cross-modal feature extraction, edge-guided feature fusion and salience map prediction. Firstly, cross-modal feature extraction module is explored to capture and aggregate united information and intersecting information in each local region of the RGB and thermal images and extracts feature-wise information. Then, considering that edge information is very helpful in refining the edge details of significant areas, edge-guided feature fusion module is explored to enhance the edge features of salient region. Moreover, the layer-bylayer decoding structure is designed to integrate the multi-level features and generate the prediction of salient maps. Compared with the state-of-the-art algorithms on all three datasets, our proposed method has the best performance. Our source codes have been available at https://github.com/0shuihan/EGFF-Net.