D2ADA: Dynamic Density-aware Active Domain Adaptation for Semantic Segmentation
This work addresses the problem of reducing annotation costs for semantic segmentation in domain adaptation, offering an incremental improvement over existing methods.
The paper tackles the trade-off between model performance and annotation cost in domain adaptation for semantic segmentation by proposing D2ADA, a dynamic density-aware active domain adaptation framework that selects informative target domain samples for labeling; it achieves comparable results to full supervision with less than 5% target annotations on benchmarks like GTA5 -> Cityscapes and SYNTHIA -> Cityscapes.
In the field of domain adaptation, a trade-off exists between the model performance and the number of target domain annotations. Active learning, maximizing model performance with few informative labeled data, comes in handy for such a scenario. In this work, we present D2ADA, a general active domain adaptation framework for semantic segmentation. To adapt the model to the target domain with minimum queried labels, we propose acquiring labels of the samples with high probability density in the target domain yet with low probability density in the source domain, complementary to the existing source domain labeled data. To further facilitate labeling efficiency, we design a dynamic scheduling policy to adjust the labeling budgets between domain exploration and model uncertainty over time. Extensive experiments show that our method outperforms existing active learning and domain adaptation baselines on two benchmarks, GTA5 -> Cityscapes and SYNTHIA -> Cityscapes. With less than 5% target domain annotations, our method reaches comparable results with that of full supervision. Our code is publicly available at https://github.com/tsunghan-wu/D2ADA.