Title: Dynamic Knowledge Distillation with A Single Stream Structure for RGB-DSalient Object Detection

Abstract: RGB-D salient object detection(SOD) demonstrates its superiority on detecting in complex environments due to the additional depth information introduced in the data. Inevitably, an independent stream is introduced to extract features from depth images, leading to extra computation and parameters. This methodology which sacrifices the model size to improve the detection accuracy may impede the practical application of SOD problems. To tackle this dilemma, we propose a dynamic distillation method along with a lightweight framework, which significantly reduces the parameters. This method considers the factors of both teacher and student performance within the training stage and dynamically assigns the distillation weight instead of applying a fixed weight on the student model. Extensive experiments are conducted on five public datasets to demonstrate that our method can achieve competitive performance compared to 10 prior methods through a 78.2MB lightweight structure.