Datasets:

Project-AgML
/

greenhouse_leafy_segmentation

dataset_info:
  features:
    - name: image
      dtype: image
    - name: mask
      dtype: image
  splits:
    - name: train
      num_bytes: 91123930
      num_examples: 3348
  download_size: 86330172
  dataset_size: 91123930
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
license: cc-by-4.0
task_categories:
  - image-segmentation
size_categories:
  - 1K<n<10K

Greenhouse Leafy Segmentation

A dataset for semantic segmentation of plants growing in a greenhouse hydroponic environment. The dataset contains 3,348 images with pixel-level mask annotations.

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

Citation

@article{cardenasgallegos2025reliable,
  title={Reliable plant segmentation under variable greenhouse illumination conditions},
  author={Cardenas-Gallegos, Jonathan S. and Severns, Paul M. and Klimeš, Pavel and N. Lacerda, Lorena and Peduzzi, Alicia and Soranz Ferrarezi, Rhuanito},
  journal={Computers and Electronics in Agriculture},
  volume={229},
  pages={109711},
  year={2025},
  publisher={Elsevier}
}

Jonathan Cárdenas. (2024). JSCG95/leafy-segmentation_CEA: Data availability release (Version V1.0.0) [Computer software]. Zenodo. https://doi.org/10.5281/zenodo.11110409

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