Bilateral Knowledge Distillation for Unsupervised Domain Adaptation of Semantic Segmentation

Unsupervised domain adaptation (UDA) aims to learn domain-invariant representations between the labeled source domain and the unlabeled target domain. Existing selftraining-based UDA methods use ground truth and pseudolabels to supervise source data and target data respectively. However, strong supervision in the source domain and pseudolabel noise in the target domain lead to some problems, such as biased predictions and over-fitting. To tackle these issues, we propose a novel Bilateral Knowledge Distillation (BKD) framework for UDA in semantic segmentation, which adopts different knowledge distillation strategies depending on the domain. Specifically, we first introduce a Source-Flow Distillation (SD) to smooth the labels of source images, which weakens the supervision in the source domain. Meanwhile, a Target-Flow Distillation (TD) is designed to extract the interclass knowledge in the probability map output from the teacher model, which alleviates the influence of pseudo-label noise in the target domain. Considering the class imbalance in semantic segmentation, we further propose an Image-Wise Hard Pixel Mining (HPM) to address this issue without estimating class frequency in the unlabeled target domain. The effectiveness of our framework against existing state-of-the-art methods is demonstrated by extensive experiments on two benchmarks: GTA5-to-Cityscapes and SYNTHIA-to-Cityscapes.