DRIV100 (Diverse Roadscenes from Internet Videos 100)

DRIV100 is a dataset for benchmarking unsupervised domain adaptation techniques on in-the-wild road-scene videos. The dataset consists of pixel-level annotations for 100 videos selected from YouTube to cover diverse scenes/domains. We provide multiple manually labeled ground-truth frames for each video, enabling a thorough evaluation of video-level domain adaptation where each video independently serves as the target domain. Please refer to our paper for more details, and please cite it if you use the DRIV100 dataset in your academic research: Haruya Sakashita, Christoph Flothow, Noriko Takemura, and Yusuke Sugano. "DRIV100: In-The-Wild Multi-Domain Dataset and Evaluation for Real-World Domain Adaptation of Semantic Segmentation." <em>arXiv preprint arXiv:2102.00150</em> (2021). https://arxiv.org/abs/2102.00150