README.md

---
dataset_info:
  features:
  - name: height
    dtype: int64
  - name: width
    dtype: int64
  - name: fold
    dtype: string
  - name: raster_name
    dtype: string
  - name: location
    dtype: string
  - name: image
    dtype: image
  - name: tile_name
    dtype: string
  - name: annotations
    struct:
    - name: bbox
      sequence:
        sequence: float64
    - name: segmentation
      list:
      - name: size
        sequence: int64
      - name: counts
        dtype: string
    - name: area
      sequence: float64
    - name: iscrowd
      sequence: int64
    - name: category
      sequence: string
  - name: tile_metadata
    struct:
    - name: crs
      dtype: string
    - name: transform
      sequence: float64
    - name: bounds
      sequence: float64
    - name: width
      dtype: int64
    - name: height
      dtype: int64
    - name: count
      dtype: int64
    - name: dtypes
      sequence: string
    - name: nodata
      dtype: float64
  splits:
  - name: validation
    num_bytes: 797545826
    num_examples: 331
  - name: test
    num_bytes: 668515932
    num_examples: 311
  download_size: 1461836536
  dataset_size: 1466061758
configs:
- config_name: default
  data_files:
  - split: validation
    path: data/validation-*
  - split: test
    path: data/test-*
license: cc-by-4.0
---

## Preprocessed Detectree2 Dataset

This version of the Detectree2 dataset has been pre-processed for use in our SelvaBox paper.

### Citation

If you use this dataset, please cite the original paper too:

```bibtex
@misc{ball_2023_8136161,
  author    = {Ball, James and Jackson, Tobias and Hickman, Sebastian and Koay, Xian Jing},
  title     = {Crown Data for ``Accurate delineation of individual tree crowns in tropical forests from aerial RGB imagery using Mask R-CNN''},
  month     = apr,
  year      = 2023,
  publisher = {Zenodo},
  version   = {v2},
  doi       = {10.5281/zenodo.8136161},
  url       = {https://doi.org/10.5281/zenodo.8136161},
}