# Mars-Bench Dust Devil Detection, MRO CTX (YOLO/COCO/VOC boxes)

## Facts

- Provider: Mirali Purohit, Hannah Kerner et al. (Mars-Bench, Arizona State University)
- License: CC-BY-4.0 (https://arxiv.org/html/2510.24010v1)
- Size: 14 files
- Format: Parquet (in sampled files)
- Stored at: Hugging Face, https://huggingface.co/datasets/Mirali33/mb-dust_devil_det
- Access: public, readable without an account
- DOI: https://doi.org/10.48550/arXiv.2510.24010
- Category: Astronomy & Space
- Source: https://huggingface.co/datasets/Mirali33/mb-dust_devil_det
- Page: https://scigantic.com/archive/996557d2-4e9a-555b-87e5-9f76b325bcd3
- Keywords: Astronomy & Space, Parquet, Hugging Face

## About

Object detection of active dust devils, the convective vortices that lift dust from the Martian surface, in roughly 750 x 750 pixel MRO Context Camera (CTX) images from the dust devil dataset of Guo et al. (2024). One class (dustdevil); splits are 1,404 training, 201 validation and 402 test images, with 1 to 50 percent training fractions. Boxes ship in YOLO format (data_yolo.yaml) and in COCO and Pascal VOC formats. Dust devil counts track near-surface atmospheric activity and dust lifting, a key part of the Martian dust cycle. Parquet shards under data/ (train in two shards); load with datasets.load_dataset("Mirali33/mb-dust_devil_det"). About 2.9 GB. Part of Mars-Bench (Purohit et al., NeurIPS 2025 Datasets and Benchmarks track, arXiv:2510.24010), 20 standardized Mars datasets released by the authors under CC BY 4.0; cite Mars-Bench and Guo et al. (2024).

## Load the data

```python
from datasets import load_dataset
ds = load_dataset("Mirali33/mb-dust_devil_det")
print(ds)
```

## Data structure

File types in sampled files: Parquet

Example file: `data/0.01x_partition-00000-of-00001.parquet`

## Open it

- In a hosted notebook with the data mounted: https://scigantic.com/archive/996557d2-4e9a-555b-87e5-9f76b325bcd3
- Full record as JSON (schema card included): https://api.scigantic.com/api/archives/996557d2-4e9a-555b-87e5-9f76b325bcd3
- Over MCP: `get_schema_card` with id `996557d2-4e9a-555b-87e5-9f76b325bcd3` at https://api.scigantic.com/mcp

## Related topics

- [Parquet datasets](https://scigantic.com/topics/parquet)
- [Astronomy & Space datasets](https://scigantic.com/topics/astronomy-and-space)
