Datasets:
dataset_info:
features:
- name: image
dtype: image
- name: mask
dtype: image
- name: objects
struct:
- name: bbox
list:
list: float64
- name: categories
list:
class_label:
names:
'0': cob
'1': ruler
- name: image_set
dtype: string
- name: population
dtype: string
- name: annotation_set
dtype: string
- name: split
dtype: string
splits:
- name: train
num_bytes: 710351925
num_examples: 1248
download_size: 710036641
dataset_size: 710351925
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
license: cc-by-nc-4.0
task_categories:
- object-detection
size_categories:
- 1K<n<10K
Deepcob Maize Cob Segmentation
Annotated images from DeepCob, a Mask R-CNN pipeline for maize cob phenotyping built on photographs of native maize landraces from the Peruvian maize genebank. Cobs laid out next to a ruler are outlined with polygons, given here as a mask (background, cob, ruler) and matching boxes. The 1,248 images come from the record's annotated training and validation sets, drawn from the older (ImgOld) and newer (ImgNew) genebank photo series and a separate cross-population set (ImgCross); the genebank population of origin is kept where the record lists it. The dataset contains 1,248 images with 8,919 bounding box annotations across 2 categories.
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
@article{kienbaum2021deepcob,
title={DeepCob: precise and high-throughput analysis of maize cob geometry using deep learning with an application in genebank phenomics},
author={Kienbaum, Lydia and Correa Abondano, Miguel and Blas, Raul and Schmid, Karl},
journal={Plant Methods},
volume={17},
pages={91},
year={2021},
publisher={BioMed Central}
}
The dataset itself can be cited as:
Kienbaum, L., Abondano, M. C., Blas, R., & Schmid, K. (2021). DeepCob: Precise and high-throughput analysis of maize cob geometry using deep learning with an application in genebank phenomics [Graphic]. Zenodo. https://doi.org/10.5281/ZENODO.4587304
This dataset was reformatted from its original format to match HuggingFace standards.