NiftyNet allows users create new network, and choose to use the network by specifying command line argument. To fully utilise this feature, a customised network should be prepared in the following steps:
Create a new network file, e.g., my_network_collection/new_net.py and make
sure my_network_collection folder on the system $PYTHONPATH, these will
ensure the network module (Python class) could be imported by the Python
interpreter.
Create a new class, e.g. NewNet in new_net.py by inheriting the
BaseNet class from niftynet.network.base_net. niftynet.network.toynet
is a minimal working example of a fully convolutional network, can be a
starting point for NewNet.
In the NewNet class, implement __init__() function for network property
initialisations, and implement layer_op() for network connections.
The network properties can be used to specify the number of channels, kernel dilation factors, as well as sub-network components of the network.
An example of sub-networks composition is presented in Simulator Gan.
The layer operation function layer_op() should specify how the input
tensors are connected to network layers. For basic building blocks, using
the ones in niftynet/layer/ are recommended. as the layers are implemented
in a modular design (convenient for parameter sharing) and can handle 2D,
2.5D and 3D cases in a unified manner whenever possible.
Finally training the network could be done by specifying the newly
implemented network in the command line argument --name my_network_collection.new_net.NewNet
(my_network_collection.new_net refer to the new_net.py file, and NewNet
is the class name with in new_net.py)
e.g., training new network with segmentation application using pip installed NiftyNet:
net_segment train -c /path/to/customised_config \
--name my_network_collection.new_net.NewNet
or using NiftyNet cloned from GitHub or CMICLab:
python net_segment.py train -c /path/to/customised_config \
--name my_network_collection.new_net.NewNet
Please consider submitting the design to our model zoo.
This section summarises the implemented network models in network.
All networks can be applied in 2D, 2.5D and 3D configurations and are reimplemented from their original presentation with their default parameters.
Reimplementation of
Çiçek, Ö., Abdulkadir, A., Lienkamp, S. S., Brox, T., and Ronneberger, O. (2016). 3D U-net: Learning dense volumetric segmentation from sparse annotation, In MICCAI 2016
- Image size - 4 should be divisible by 8
- Label size should be more than 88
- border is 44
Reimplementation of
Milletari, F., Navab, N., & Ahmadi, S. A. (2016). V-net: Fully convolutional neural networks for volumetric medical image segmentation, In 3DV 2016
- Image size should be divisible by 8
Implementation of
Fidon, L., Li, W., Garcia-Peraza-Herrera, L.C., Ekanayake, J., Kitchen, N., Ourselin, S., Vercauteren, T. (2017). Scalable convolutional networks for brain tumour segmentation. In MICCAI 2017
- More than one modality should be used
Implementation of
Li W, Wang G, Fidon L, Ourselin S, Cardoso M J, Vercauteren T, (2017). On the compactness, efficiency, and representation of 3D convolutional networks: Brain parcellation as a pretext task, In IPMI 2017
- Image size should be divisible by 8
Reimplementation of
Kamnitsas, K., Ledig, C., Newcombe, V. F., Simpson, J. P., Kane, A. D., Menon, D. K., Rueckert, D., Glocker, B. (2017). Efficient multi-scale 3D CNN with fully connected CRF for accurate brain lesion segmentation, MedIA 36, 61-78
- The downsampling factor (
d_factor) should be odd - Label size = [(image_size / d_ factor) - 16 ]*d_factor
- Image size should be divisible by d_factor
Example of appropriate configuration for training:
image spatial window size = 57, label spatial window size = 9, d_ factor = 3
and for inference:
image spatial window size = 105, label spatial window size = 57, d_ factor = 3
Implementation of
Fidon, L. et. al. (2017) Generalised Wasserstein Dice Score for Imbalanced Multi-class Segmentation using Holistic Convolutional Networks. MICCAI 2017 (BrainLes)