MC-Net: A Multi-Modal Cross-Scale Feature Interaction Learning Method for RGB-T Salient Object Detection
Qi Qi · IEICE Transactions on Fundamentals of Electronics Communications and Computer Sciences · 2025
This letter proposes the MC-Net, which aims to improve salient object detection accuracy in complex environments by integrating multimodal information. By introducing cross-attention and cross-scale mechanism, the method facilitates the interactive learning between different feature representations. Experimental results show that MC-Net can make a significant breakthrough in detection performance.