Sta8is/Nuscenes_depthv2

nuScenes depth labels (Depth Anything V2)

Depth labels for the front camera (CAM_FRONT) of nuScenes, used by DINO-Foresight to evaluate the future prediction of depth on nuScenes. The maps are generated with Depth Anything V2 (ViT-L), in the same way as the Cityscapes depth labels of DINO-Foresight. The dataset contains only the labels, not the nuScenes images.

Content

Labels for 4,164 front-camera frames (669 keyframes in samples and 3,495 intermediate frames in sweeps): 450 frames of the validation scenes and 3,714 frames of the training scenes. This includes the evaluation frame (frame 28 of the scene) of every one of the 700 training and 150 validation scenes.

Each label is an 8-bit grayscale JPEG of size 1600x900, with the same file name and relative path as the corresponding nuScenes image:

nuscenes_depthv2
│
├───samples
│   └───CAM_FRONT          # <scene>__CAM_FRONT__<timestamp>.jpg
├───sweeps
│   └───CAM_FRONT

Download

hf download Sta8is/Nuscenes_depthv2 --repo-type dataset --local-dir NuScenes_depthv2

Place the folder next to the nuScenes folder (NuScenes and NuScenes_depthv2), or pass its location with --depth_path in DINO-Foresight. See the DINO-Foresight repository for the nuScenes setup.

License

The labels are derived from nuScenes images. The nuScenes terms of use (CC BY-NC-SA 4.0, non-commercial use) apply.

Citation

If you use this dataset, please cite DINO-Foresight and nuScenes.

@inproceedings{karypidis2025dinoforesight,
title={{DINO}-Foresight: Looking into the Future with {DINO}},
author={Efstathios Karypidis and Ioannis Kakogeorgiou and Spyros Gidaris and Nikos Komodakis},
booktitle={The Thirty-ninth Annual Conference on Neural Information Processing Systems},
year={2025},
url={https://arxiv.org/abs/2412.11673}
}