This is the source code of "Automated Recognition of Right Ventricular Dilatation and Prognosis Correlation", implemented with Keras and Tensorflow.
- All pretrained classification models were stored under the "model" folder.
- run.ipynb: notebook of how you load and use the models or train your own model.
- For model architecture and training, please refer to the "src" folder.
Application was shown in survival_analysis.ipynb.
The segmentation results came from our pretrained model. Adding both segmentation channels of left and right ventricle, the input arrays become larger.
We trained the A4C-seg model with these reproduced data, which is visualized and shown as the following gif.
