FundusNet: a deep learning approach for identifying novel endophenotypes for neurodegenerative and eye diseases from fundus images
Hu, W., Li, K., Gagnon, J., Wang, Y., Raney, T., Chen, J., Chen, Y., Okunuki, Y., Chen, W., & Zhang, B. (2025). FundusNet: A Deep-Learning Approach for Fast Diagnosis of Neurodegenerative and Eye Diseases Using Fundus Images. Bioengineering, 12(1), 57. https://doi.org/10.3390/bioengineering12010057
- git clone the repo
- Execute either shgender.sh or shage.sh to run individual CNN or ViT models:
a. This process will split the image dataset into training and testing sets, train the CNN/ViT models on the training data, and evaluate them on the test data.
b. Users must provide the following inputs:
'name of csv_file (string)': Path to the CSV file containing annotations.
'root_dir (string)': Directory containing all images. - Combine the results using majority voting for ensemble prediction.