DeLiVER: Delivering Arbitrary-Modal Semantic Segmentation

[CVPR 2023] Official Dataset Repository for the DeLiVER benchmark.

• Project Page: https://jamycheung.github.io/DELIVER.html
• GitHub Repository: https://github.com/InSAI-Lab/DELIVER
• Paper: arXiv:2303.01480
• Model Checkpoints: InSAI-Lab/CMNeXt

---

Dataset Description

To conduct arbitrary-modal semantic segmentation, we created the DeLiVER benchmark, covering Depth, LiDAR, multiple Views, Events, and RGB. It includes 4 adverse weather conditions (cloudy, foggy, night-time, sunny) and 5 sensor failure corner cases:

• MB: Motion Blur
• OE: Over-Exposure
• UE: Under-Exposure
• LJ: LiDAR-Jitter
• EL: Event Low-resolution

Each sample has 6 views, 4 modalities, and 2 annotation types (semantic and instance).

File Structure

The dataset archive DELIVER.tar.gz (~12.2 GB) unpacks into:

DELIVER/
├── depth/
│   ├── cloud/ (test, train, val)
│   ├── fog/
│   ├── night/
│   ├── rain/
│   └── sun/
├── event/
├── hha/
├── img/
├── lidar/
└── semantic/

Quick Start & Download

Using the Hugging Face CLI:

# Download dataset archive
hf download InSAI-Lab/DELIVER DELIVER.tar.gz --type dataset --local-dir ./data

# Extract archive
cd ./data
tar -zxvf DELIVER.tar.gz

Using Python (huggingface_hub):

from huggingface_hub import hf_hub_download

dataset_path = hf_hub_download(
    repo_id="InSAI-Lab/DELIVER",
    repo_type="dataset",
    filename="DELIVER.tar.gz",
    local_dir="./data"
)

Citation

@inproceedings{zhang2023delivering,
  title={Delivering Arbitrary-Modal Semantic Segmentation},
  author={Zhang, Jiaming and Liu, Ruiping and Shi, Hao and Yang, Kailun and Rei{\ss}, Simon and Peng, Kunyu and Fu, Haodong and Wang, Kaiwei and Stiefelhagen, Rainer},
  booktitle={CVPR},
  year={2023}
}

@article{zhang2023cmx,
  title={CMX: Cross-modal fusion for RGB-X semantic segmentation with transformers},
  author={Zhang, Jiaming and Liu, Huayao and Yang, Kailun and Hu, Xinxin and Liu, Ruiping and Stiefelhagen, Rainer},
  journal={IEEE Transactions on Intelligent Transportation Systems},
  year={2023}
}

License

This dataset is released under the Apache-2.0 License.