Datasets:

Project-AgML
/

QuinceSet_detection

dataset_info:
  features:
    - name: image
      dtype: image
    - name: objects
      struct:
        - name: bbox
          list:
            list: float64
        - name: categories
          list:
            class_label:
              names:
                '0': Ripe quince
                '1': Unripe quince
  splits:
    - name: train
      num_bytes: 4473166727
      num_examples: 1515
  download_size: 5812591565
  dataset_size: 4473166727
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
license: cc-by-4.0
task_categories:
  - object-detection
size_categories:
  - 1K<n<10K

QuinceSet Detection

A dataset for detection of quince fruit. The dataset contains 1,515 images with 17,171 bounding box annotations across 2 categories.

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

Citation

@article{kaufmane2022quinceset,
  title={QuinceSet: Dataset of annotated Japanese quince images for object detection},
  author={Kaufmane, Ed{\=\i}te and Sudars, Kaspars and Namat{\=e}vs, Ivars and Kalni{\c{n}}a, Ieva and Judvaitis, J{\=a}nis and Bala{\v{s}}s, Rihards and Strauti{\c{n}}a, Sarm{\=\i}te},
  journal={Data in Brief},
  volume={42},
  pages={108332},
  year={2022},
  publisher={Elsevier}
}

Kaufmane, E., Sudars, K., Namatēvs, I., Kalniņa, I., Judvaitis, J., Balašs, R.& Strautiņa, S. (2022). QuinceSet: Dataset of Annotated Japanese Quince Images for Object Detection. Data in Brief. https://doi.org/10.5281/zenodo.6402251

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