Datasets:

Project-AgML
/

roseleafset_disease_classification

dataset_info:
  - config_name: augmented
    features:
      - name: image
        dtype: image
      - name: label
        dtype:
          class_label:
            names:
              '0': Black Spot
              '1': Dry Leaf
              '2': Healthy Leaf
              '3': Leaf Holes
    splits:
      - name: train
        num_bytes: 400562967
        num_examples: 10000
    download_size: 400801970
    dataset_size: 400562967
  - config_name: raw
    features:
      - name: image
        dtype: image
      - name: label
        dtype:
          class_label:
            names:
              '0': Black Spot
              '1': Dry Leaf
              '2': Healthy Leaf
              '3': Leaf Holes
    splits:
      - name: train
        num_bytes: 5570900810
        num_examples: 3113
    download_size: 5571111871
    dataset_size: 5570900810
configs:
  - config_name: augmented
    data_files:
      - split: train
        path: augmented/train-*
  - config_name: raw
    data_files:
      - split: train
        path: raw/train-*
    default: true
license: cc-by-4.0
task_categories:
  - image-classification
size_categories:
  - 1K<n<10K

Roseleafset Disease Classification

RoseLeafSet is a set of close-up smartphone photos of rose leaves collected in rose nurseries around Savar, Dhaka, Bangladesh, between October 30 and November 6, 2024, labelled into four foliar conditions: black spot, dry leaf, healthy leaf and leaf holes. Images were taken with a Vivo IQOO Z9x phone at 3264 x 2448 pixels. The default raw config holds the 3,113 original photos (class-imbalanced: 1,288 black spot, 324 dry leaf, 818 healthy, 683 leaf holes); the augmented config holds the authors' class-balanced augmented set of 2,500 images per class (10,000 total), derived from contrast-enhanced, padded versions of the originals. The dataset contains raw and augmented versions.
The raw dataset contains 3,113 images.
Images per class:

Black Spot: 1,288
Dry Leaf: 324
Healthy Leaf: 818
Leaf Holes: 683

The augmented dataset contains 10,000 images.
Images per class:

Black Spot: 2,500
Dry Leaf: 2,500
Healthy Leaf: 2,500
Leaf Holes: 2,500

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

Citation

<!-- TODO: add BibTeX citation -->

The dataset itself can be cited as:

Jinia, J. T., Rahman, M. S., Mojumdar, M. U., & Bijoy, M. H. I. (2025). RoseLeafSet: Real-World Leaf Image Dataset for AI-Based Agricultural Solutions [Dataset]. Mendeley Data. https://doi.org/10.17632/9G668BFHY5.3

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