This page 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: image_ size = 57, label_ size = 9, d_ factor = 3
- Create a
niftynet/network/new_net.py
inheritingBaseNet
fromniftynet.layer.base_net
- Implement
layer_op()
function using the building blocks inniftynet/layer/
or creating new layers - Import
niftynet.network.new_net
to theNetFactory
class inniftynet/__init__.py
- Train the network with
python net_segment.py train -c /path/to/customised_config