CC-Cruiser is an intelligence agent involving three functional networks: “identification networks” for screening for CC in populations, “evaluation networks” for risk stratification among CC patients and “strategist networks” to assist in treatment decisions by ophthalmologists.
Please feel free to contact us for any questions or comments: Erping Long, E-mail: [email protected].
For auto-cutting, the "cut.m" is the startup file and could be executed in MATLAB. The representative samples before and after auto-cutting is presented.
All codes for deep-learning convolutional neural networks were executed in the Caffe (Convolutional Architecture for Fast Feature Embedding) framework with Ubuntu 14.04 64bit + CUDA (Compute Unified Device Architecture).
The /DL-Source code/createdata contains the dataset for one-time training and testing. The training and testing records are saved as test.txt and train.txt. The script "create_imagenet.sh" is used to generate the dataset. The file "make_imagenet_mean.sh" is used to generate the mean file for dataset accordingly.
The file "train.sh" in /DL-Source code/myself is used for network training.
The file "test.sh" in /DL-Source code/myself is used for testing.
The "yunxing_cnn.py" in /DL-Source code/python_script could be used to test and have the evaluation indices.