---
dataset_info:
  features:
  - name: image
    dtype: image
  - name: label
    dtype:
      class_label:
        names:
          '0': healthy
          '1': flood_stress
          '2': bbwv2_infected
  - name: capture_date
    dtype: string
  - name: source_photo
    dtype: string
  splits:
  - name: train
    num_bytes: 89717136
    num_examples: 4102
  download_size: 89670846
  dataset_size: 89717136
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
license: cc-by-4.0
task_categories:
- image-classification
size_categories:
- 1K<n<10K
---
# Chili Pepper Virus Classification

4,102 chili pepper leaf crops (256x256) released with a study on transferring plant disease classifiers from laboratory to field images, labelled healthy, flooding stress or broad bean wilt virus 2 (BBWV2) infection. The infected plants came from the Plant Molecular Immunology Lab at Seoul National University. The capture date and the source photo each crop was cut from are kept as columns. The dataset contains 4,102 images across 3 classes: healthy, flood_stress, bbwv2_infected.  
Images per class:
- healthy: 1,325
- flood_stress: 1,642
- bbwv2_infected: 1,135

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

## Citation

```bibtex
@article{jeon2026bridging,
  title={Bridging the Lab-to-Field gap in plant disease diagnosis through unsupervised domain adaptation enhanced by background recomposition},
  author={Jeon, Woosang and Kim, Taehyeong and Choi, Sanghyeok and Yang, Kyuseok and Kim, Seong-Yeop and Song, Mungyeong},
  journal={Ecological Informatics},
  volume={93},
  pages={103579},
  year={2026},
  publisher={Elsevier}
}
```


The dataset itself can be cited as:

Kim, S.-Y., Song, M., Jeon, W., Kim, T., Choi, S., &amp; Yang, K. (2025). *Chili Pepper BBWV2 &amp; Overwatering Stress Leaf Image Dataset* [Graphic]. Zenodo. https://doi.org/10.5281/ZENODO.17901489


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