Collaborative Learning for Multi-Source Domain Adaptative Object Detection

Multi-Source domain adaptation is a more general scenario than single-source domain adaptation, however, it encounters challenges arising from the problem of model effect degradation caused by distribution differences between multiple source domains. To address this issue, we propose a novel approach, namely collaborative learning(CL). CL can retain public intra-domain knowledge and improve the generalization ability of the model. Extensive experiments on different adaptation scenarios demonstrate the effectiveness of the proposed model.