Setup Instructions

Model Weights

1. Download SRCNN weights and place it in server/label_scale_bar_detector/OCR/SRCNN-pytorch/weights/.
2. Download Darknet weights and place it in server/label_scale_bar_detector/localizer/darknet/backup.
3. Download Mask RCNN weights and place it in server/Mask_RCNN/logs/tem.

Installation

Run conda env create -f environment.yml. Note: Installation using this yml file has been tested only on MACOSX.

Running the app

1. export CORS_DEBUG="True"
2. cd client && npm start.
3. cd server && uvicorn main:app --reload.

The following is an illustration of the results displayed when a sample TEM image is uploaded.

Acknowledgements

1. https://github.com/AlexeyAB/darknet
2. https://github.com/yjn870/SRCNN-pytorch
3. https://github.com/matterport/Mask_RCNN

Citation

If you use this code, please cite the following manuscript:

@misc{subramanian2021dataset,
      title={Dataset of gold nanoparticle sizes and morphologies extracted from literature-mined microscopy images}, 
      author={Akshay Subramanian and Kevin Cruse and Amalie Trewartha and Xingzhi Wang and Paul Alivisatos and Gerbrand Ceder},
      year={2021},
      eprint={2112.01689},
      archivePrefix={arXiv},
      primaryClass={cond-mat.mtrl-sci}
}