Datasets:

Project-AgML
/

coffee_rust_multispec_classification

dataset_info:
  features:
    - name: rgb
      dtype: image
    - name: bands
      dtype: binary
    - name: bands_shape
      list: int64
    - name: bands_dtype
      dtype: string
    - name: label
      dtype:
        class_label:
          names:
            '0': NoRust
            '1': Rust
  splits:
    - name: train
      num_bytes: 336062712
      num_examples: 1120
  download_size: 93392012
  dataset_size: 336062712
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
license: cc0-1.0
task_categories:
  - image-classification
size_categories:
  - 1K<n<10K

Coffee Rust Multispec Classification

A dataset for image classification of Coffee Rust Multispec Classification. The dataset contains 1,120 images across 2 classes: NoRust, Rust.
Images per class:

NoRust: 273
Rust: 847

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

Citation

@article{arocatrujillo2025colombian,
  title={Colombian coffee tree leaves multispectral images dataset},
  author={Aroca-Trujillo, Jorge Luis and Perez-Ruiz, Alexander},
  journal={Data in Brief},
  volume={59},
  pages={111421},
  year={2025},
  publisher={Elsevier}
}

https://www.kaggle.com/datasets/jorgearoca/coffee-rust

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