Defoliation Dataset
Published by Zichen Zhang, Sami Khanal, and Christopher Stewart

This dataset contains RGB images of crop fields growing soybean.  The images have been labeled according to the
degree of defoliation present in the image.  Additional, labels can be added later.

Importantly, this dataset is structured to support the OSU ICICLE Instistute Data and Model Commons.  The Data 
Commons stores datasets and their metadata together to support download and sharing between researchers.  The
Model Commons allows interested parties to search for models that perform well on representative datasets for
use in their domain.

Organization:

-- Subdirectory: RawData

This subdirectory contains the collected images and labels.  Here, these have been compressed into a Zip file.

-- Subdirectory: Datasheet

This subdirectory contains a YAML file describing salient qualitative characteristics of the dataset.  See 
the Datasheets for Datasets paper for more details (https://arxiv.org/abs/1803.09010).  As additional labels
are added/identified for the dataset, more YAML files can be added here.

-- Sub Directory: Summary

This subdirectory contains a YAML file characterizing the results of models (machine learned or hand-crafted)
that have been tested using the dataset.  Again, more YAML files can be added as models grow.

-- Subdirectory: Reference

This subdirectory contains a YAML file with pointers to models that have been executed in the past.  In some
cases, the models may be saved to files in the rawmodels directory for convenience.  When possible, we prefer
to save models in Docker containers for easy reuse and fewer dependency concerns.