基于深度神经网络对眼底图像进行侧向性分类.zip

# Fundus Laterality # ### Laterality Classification of Fundus Images using Interpretable Deep Neural Networks [[original paper]()] [[cite]()]### ## Package Dependency ## matplotlib==1.3.1 numpy==1.14.3 scikit_image==0.10.1 pandas==0.21.1 scipy==0.19.1 Keras==2.1.6 Pillow==5.1.0 skimage==0.0 scikit_learn==0.19.1 ## Running the code ## At the home directory, install required packages with ``` pip install requirements.txt``` Run inferece with ``` python inference.py --test_dir=<folder_path> --gpu_index=<gpu_index>``` For instance, run inferece for the sample dataset ([IDRiD dataset](https://idrid.grand-challenge.org/)) with ``` python inference.py --test_dir=IDRiD_imgs/ --gpu_index=0``` IDRiD_NAR_013.jpg is wrongly classified in the sample dataset. ## File Hierarchy ## . ├── IDRiD_imgs ├── inference.py ├── model │   └── relu_softmax_weight_epoch49.h5 ├── network.py ├── README.md ├── requirements.txt ├── results.csv └── utils.py IDRiD_imgs: sample fundus images inference.py: inference file to run model/relu_softmax_weight_epoch49.h5: weight file network.py: network file in keras style results.csv: results of laterality with columns of filename and the results (R or L) utils.py: miscellaneous helper functions ## LICENSE ## This is under the MIT License Copyright (c) 2018 Vuno Inc. (www.vuno.co)