This code was used to train the AxoNet convolutional neural network. To watch the model train on the dataset, run main.py and sit back.
We packaged AxoNet into a user-friendly Fiji plugin. Quick-use instructions can be found here.
Python 3, matplotlib, numpy, skimage, PIL, scipy, keras, and tensorflow.
We recommend Anaconda for dependency management.
This work is licensed under a Apache 2.0 license.
Users at for-profit commercial organizations are requested to provide a donation, in an amount of their choice, to support the wonderful work of the National Glaucoma Research Program of the BrightFocus Foundation.
Users in the non-profit sector, i.e. at academic, foundation and governmental organizations, are requested to consider a similar donation.
If you publish papers using this software, please cite this paper.
This tool was developed by the Ethier Lab at the Georgia Institute of Technology.
Run main.py to download the rat optic nerve image and annotation dataset.
The results from our counting method comparisons are located in "final results- comparison with both nhp and rat.xlsx" inside the data folder.