# PApple_RGB-D-Size: Fuji Apple Tree RGB-D Images with Modal and Amodal Fruit Masks and Diameter Ground Truth, 3,925 Images (Universitat de Lleida, AgML Mirror)

## Facts

- Provider: Jordi Gené Mola, Mar Ferrer Ferrer, Eduard Gregorio López, Joan Ramon Rosell Polo, Verónica Vilaplana and Javier Ruiz Hidalgo (Universitat de Lleida and Universitat Politècnica de Catalunya); Hugging Face mirror by Project-AgML
- License: CC-BY-4.0 (https://zenodo.org/records/7260694)
- Size: 121 files
- Format: Parquet (in sampled files)
- Stored at: Hugging Face, https://huggingface.co/datasets/Project-AgML/papple_rgbd_amodal_segmentation
- Access: public, readable without an account
- DOI: https://doi.org/10.5281/zenodo.7260694
- Category: Agriculture & Environment
- Source: https://huggingface.co/datasets/Project-AgML/papple_rgbd_amodal_segmentation
- Page: https://scigantic.com/archive/876f2cb5-6491-5aa3-bec3-ad4fdcdf74df
- Keywords: Agriculture & Environment, image segmentation, Parquet, Hugging Face

## About

PApple_RGB-D-Size holds 3,925 RGB-D images (1300 x 1300) of Fuji apple trees at growth stages BBCH77 and BBCH85, built to estimate the size of on-tree apples that are partly hidden by leaves and other fruit. Each apple has a modal (visible) and an amodal (complete) segmentation outline, matching boxes, an apple ID and fruit diameter ground truth. Depth maps from photogrammetry are kept as the original float64 arrays, with a focal length of 5805.34 pixels for every image, and the source train, validation and test split is preserved; 15,335 apple boxes in all. The data trained the Amodal_Fruit_Sizing network from the GRAP group. Stored as Parquet shards with the image bytes and annotations in columns; load with datasets.load_dataset("Project-AgML/papple_rgbd_amodal_segmentation") or through the AgML Python library (Project-AgML, UC Davis), which reformatted the original release. License: CC BY 4.0, as stated by the original publisher; cite Jordi Gené Mola, Mar Ferrer Ferrer, Eduard Gregorio López, Joan Ramon Rosell Polo, Verónica Vilaplana and Javier Ruiz Hidalgo et al. and the DOI 10.5281/zenodo.7260694.

## Load the data

```python
from datasets import load_dataset
ds = load_dataset("Project-AgML/papple_rgbd_amodal_segmentation")
print(ds)
```

## Data structure

File types in sampled files: Parquet

Example file: `data/train-00000-of-00119.parquet`

## Open it

- In a hosted notebook with the data mounted: https://scigantic.com/archive/876f2cb5-6491-5aa3-bec3-ad4fdcdf74df
- Full record as JSON (schema card included): https://api.scigantic.com/api/archives/876f2cb5-6491-5aa3-bec3-ad4fdcdf74df
- Over MCP: `get_schema_card` with id `876f2cb5-6491-5aa3-bec3-ad4fdcdf74df` at https://api.scigantic.com/mcp

## Related topics

- [Parquet datasets](https://scigantic.com/topics/parquet)
- [Image segmentation datasets](https://scigantic.com/topics/image-segmentation)
- [Agriculture & Environment datasets](https://scigantic.com/topics/agriculture-and-environment)
