- We have used Python 3.7 and Tensorflow 2.3.
- Install Anaconda
- Follow instructions from here to create an anaconda environment from environment.yml file.
- Run generate_data.py to generate dataset.
- Run main.py to start training the model.
- After 1000 epochs training ends and to check the model working run evaluate.py file.
- To run pre-trained model create a folder with name checkpoint and after downloading checkpoint file from here, unzip it and paste it into the checkpoint folder.
- Run evaluate.py to check the working.
- Threshold value for valid plot is set to 0.5.
@misc{gangopadhyay2021apexnet,
title={APEX-Net: Automatic Plot Extractor Network},
author={Aalok Gangopadhyay and Prajwal Singh and Shanmuganathan Raman},
year={2021},
eprint={2101.06217},
archivePrefix={arXiv},
primaryClass={cs.CV}
}