A GUI based semi annotation tool to annotate chart figures according to their type (e.g. Bar chart, Line chart) from Biomedical literature (e.g. Pubmed).
- pyQt5
- numpy
- pandas
- scipy
- scikit-learn
- keras
- tensorflow
Run pip install -r requirements.txt to automatically install the dependencies in your virtual environment. Execute the viewer.py file. Tested with python 3.6 and tensorflow 1.4.
- Select open file or open csv if you have already created from the toolbar.
- Select the folder which contains the papers along with the figure charts from pubmed.
- Annotate some images by pressing right click on the scenes.
- Click the training button to train on the small subset in order to get better results.
- Continue the annotation.
