saharzafari / keras-resnet3d

Implementations of ResNets for volumetric data, including a vanilla resnet in 3D.

keras-resnet3d

Resnet 3D

A vanilla 3D extention to raghakot/keras-resnet

VoxResNet (TODO)

A keras re-implementation of VoxResNet (Hao Chen et.al) for volumetric image segmention. (Non-official)

keras-voxresnet enables volumetric image classification with keras and tensorflow/theano.

Installation

Dependencies

keras, tensorflow/theano and their corresponding dependencies.

Install with pip

$ pip install git+https://github.com/JihongJu/keras-resnet3d.git

Build from source

$ git clone https://github.com/JihongJu/keras-resnet3d.git
$ cd keras-resnet3d
$ python setup.py build

Usage

from resnet3d import Resnet3DBuilder
model = Resnet3DBuilder.build_resnet_50((96, 96, 96, 1), 20)
model.compile(optimizer='adam',
              loss='categorical_crossentropy',
              metrics=['accuracy'])
model.fit(X_train, y_train, batch_size=32)

More details see example:

• A lung-cancer-detector