SJTU-M3DV
Final project for course EE369 in SJTU, implemented in Keras+Tensorflow.
Model file 'dense.h5' is not in the file path, and can be downloaded from here.
Four different neuron networks have been used to train the model, and the final model is trained via data_train_densenet.py.
Requirements
- Python 3 (Anaconda 3.5.1 specifically)
- Tensorflow==1.14.0
- Keras==2.2.0
Code Structure
conv: a simple implementation of 3D convolutional neuron network
densenet: an implementation of 3D Densenet network, 2017 CVPR best paper
resnet3d: an implementation of 3D Resnet network
inception: an implementation of 3D Inception network
data: raw train data is not allowed to share in this repository.
data_train_cnn.py
data_train_densenet.py
data_train_resnet.py
data_train_inc.py
misc.py: basic functions
test.py: load model