[ Supplementary Material ] Discover, Hallucinate, and Adapt: Open Compound Domain Adaptation for Semantic Segmentation

A.1, we evaluate our framework on two new datasets, Synscapes and SYNTHIA, A.2, we conduct additional ablation studies on the adaptation step using four latent A.3, we analyze hyperparameter K selection. A.4, we show more qualitative results. A.5, we elaborate the implementation details. The adaptation results are summarized in the Table 1. In the main paper, we already show that the proposed domain-wise adversaries are more effective than the traditional UDA approaches.