Official implementation of "Utilizing unsupervised learning to improve sward content prediction and herbage mass estimation" paper. Published in the 29th European Grassland Federation General Meeting (2022) pp 791-793 proceedings.

train.sh contrains the commands to train the unsupervised algorithm and to use the learned weights to initialize the regression network

The irish dataset is not publicaly available The GrassClover dataset and unlabeled images can be found at https://competitions.codalab.org/competitions/21122#participate-get-data

You might want to preprocess the dataset by downscaling the unsupervised images to a lower resolution before training to speed up the image operations.

If it helped your work

@article{2022_EGF_unsupervisedclover,
  title={Utilizing unsupervised learning to improve sward content prediction and herbage mass estimation},
  author={Albert, Paul and Saadeldin, Mohamed and Narayanan, Badri and Mac Namee, Brian and Hennessy, Deirdre and O'Connor, Aisling H and O'Connor, Noel E and McGuinness, Kevin},
  journal={arXiv preprint arXiv:2204.09343},
  year={2022}
}

Albert, Paul, et al. "Utilizing unsupervised learning to improve sward content prediction and herbage mass estimation." arXiv preprint arXiv:2204.09343 (2022).