Datasets:

Project-AgML
/

DeepRootLab_root_segmentation

dataset_info:
  features:
    - name: image
      dtype: image
    - name: mask
      dtype: image
    - name: split
      dtype: string
    - name: PipeID
      dtype: string
    - name: Position
      dtype: string
    - name: timestamp
      dtype: string
  splits:
    - name: train
      num_bytes: 69854775
      num_examples: 438
  download_size: 69911174
  dataset_size: 69854775
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
license: cc-by-4.0
task_categories:
  - image-segmentation
size_categories:
  - n<1K

Deeprootlab Root Segmentation

This dataset provides real RGB images of plant root systems captured in a controlled laboratory environment for agricultural root phenotyping. It focuses on root segmentation tasks, offering high-quality imagery suitable for developing semantic segmentation models in agricultural research contexts. The dataset contains 438 images with pixel-level mask annotations.

This dataset is indexed on https://project-agml.github.io/ as part of the AgML python library.

The original train/test/val split has been preserved in the split column.

Citation

@article{han2026deep,
  title={Deep roots through time and crops: insight from five seasons at DeepRootLab},
  author={Han, Eusun and Cl{\'e}ment, Corentin and Czaban, Weronika and Smith, Abraham George and Dresb{\o}ll, Dorte Bodin and Thorup-Kristensen, Kristian},
  journal={New Phytologist},
  volume={250},
  number={4},
  pages={2670--2688},
  year={2026},
  publisher={Wiley Online Library}
}

Han, E., Clément, C., Czaban, W., Smith, A. G., Dresbøll, D. B., & Thorup-Kristensen, K. (2025). Dataset used in "Five seasons with DeepRootLab: A unique facility for easier deep root research in the field" [Dataset]. Zenodo. https://doi.org/10.5281/zenodo.15213661

This dataset was reformatted from its original format to match HuggingFace standards.