Medical imaging data are typically large in size. As a result, it can be difficult to train deep neural models on them. The activations of invertible neural networks do not have to be stored to perform backpropagation, therefore such networks can be used to save memory when handling large data volumes. We use a technique called additive coupling to obtain a memory-efficient partially-reversible image-to-image translation model. With this model, we perform a 3D super-resolution and 3D domain-adaptation task, on both paired and unpaired CT scan data. Additionally, we demonstrate experimentally that the model requires significantly less GPU memory than models without reversibility. samples as in or out-of distribution, our method achieves an AUC score of 0.99.
Chest CT Super-resolution and Domain-adaptation using Memory-efficient 3D Reversible GANs
T. van der Ouderaa, D. Worrall and B. van Ginneken
Medical Imaging with Deep Learning 2019.
A pdf file of this publication is available for personal use. Enter your e-mail address in the box below and press the button. You will receive an e-mail message with a link to the pdf file.
An email message containing a code and instructions to download the following paper has been sent to your email address.