Models

U-net

synthtorch.models.unet

holds the architecture for a 2d or 3d unet [1,2,3]

References

[1] Ronneberger, Olaf, Philipp Fischer, and Thomas Brox.

“U-net: Convolutional networks for biomedical image segmentation.” International Conference on Medical image computing and computer-assisted intervention. Springer, Cham, 2015.

[2] O. Cicek, A. Abdulkadir, S. S. Lienkamp, T. Brox, and O. Ronneberger,

“3D U-Net: Learning Dense Volumetric Segmentation from Sparse Annotation,” in Medical Image Computing and Computer-Assisted Intervention (MICCAI), 2016, pp. 424–432.

[3] C. Zhao, A. Carass, J. Lee, Y. He, and J. L. Prince, “Whole Brain Segmentation and Labeling

from CT Using Synthetic MR Images,” MLMI, vol. 10541, pp. 291–298, 2017.

Author: Jacob Reinhold (jacob.reinhold@jhu.edu)

Created on: Nov 2, 2018

Variational Autoencoder

synthtorch.models.vae

construct a variational autoencoder

Author: Jacob Reinhold (jacob.reinhold@jhu.edu)

Created on: Jan 29, 2019

N-layer CNN

synthtorch.models.nconvnet

define the class for a N layer CNN with no max pool, increase in channels, or any of that fancy stuff. This is generally used for testing purposes

Author: Jacob Reinhold (jacob.reinhold@jhu.edu)

Created on: Nov 2, 2018

DenseNet

synthtorch.models.densenet

holds the architecture for a 2d densenet [1] this model is pulled (and modified) from the pytorch repo: https://github.com/pytorch/vision/blob/master/torchvision/models/densenet.py

References

[1] Huang, Gao, et al. “Densely connected convolutional networks.”

Proceedings of the IEEE conference on computer vision and pattern recognition. 2017.

Author: Jacob Reinhold (jacob.reinhold@jhu.edu)

Created on: Apr 8, 2018