Datasets:
File size: 1,991 Bytes
3e831ab 45c5162 3e831ab 45c5162 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 | ---
dataset_info:
features:
- name: image
dtype: image
- name: objects
struct:
- name: bbox
list:
list: float64
- name: categories
list:
class_label:
names:
'0': apple_blotch_acervulus
- name: split
dtype: string
splits:
- name: train
num_bytes: 123508180
num_examples: 927
download_size: 123511401
dataset_size: 123508180
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
license: cc0-1.0
task_categories:
- object-detection
size_categories:
- n<1K
---
# Apple Blotch Detection
Detached apple leaves photographed after artificial inoculation with Diplocarpon coronariae, the fungus behind apple blotch, as part of a resistance screening of apple genetic resources (551 varieties in 2022, 80 in 2023) at the Julius Kühn Institute in Dresden-Pillnitz, Germany. Leaves were imaged on a light table with a tripod-mounted Canon DSLR, and 768 x 768 crops of the leaf surface were annotated with bounding boxes around individual acervuli (the fungus's spore-producing spots) for training a YOLOv5 counting model. Crops without symptoms are kept as negatives with no boxes; the split column keeps the authors' train/val split. The dataset contains 927 images with 4,167 bounding box annotations across 1 category.
This dataset is indexed on https://project-agml.github.io/ as part of the AgML python library.
The original train/test/val split has been preserved in the `split` column.
## Citation
```bibtex
@article{reim2024yolo,
title={YOLO-Based Phenotyping of Apple Blotch Disease (Diplocarpon coronariae) in Genetic Resources after Artificial Inoculation},
author={Reim, Stefanie and Richter, Sophie and Leonhardt, Oskar and Maß, Virginia and Wöhner, Thomas Wolfgang},
journal={Agronomy},
volume={14},
pages={1042},
year={2024},
publisher={MDPI}
}
```
*This dataset was reformatted from its original format to match HuggingFace standards.*
|