Timezone: »

Multiscale Metamorphic VAE for 3D Brain MRI Synthesis
Jaivardhan Kapoor · Jakob Macke · Christian Baumgartner

Generative modeling of 3D brain MRIs presents difficulties in achieving high visual fidelity while ensuring sufficient coverage of the data distribution. In this work, we propose to address this challenge with composable, multiscale morphological transformations in a variational autoencoder (VAE) framework. These transformations are applied to a chosen reference brain image to generate MRI volumes, equipping the model with strong anatomical inductive biases. We show substantial performance improvements in FID while retaining comparable, or superior, reconstruction quality compared to prior work based on VAEs and generative adversarial networks (GANs).

Author Information

Jaivardhan Kapoor (University of Tübingen)

I am a PhD student co-supervised by Dr. Jakob Macke and Dr. Christian Baumgartner at the University of Tübingen. My work is in developing probabilistic deep generative models primarily for neuroimaging data.

Jakob Macke (University of Tuebingen)
Christian Baumgartner (University of Tübingen)
Christian Baumgartner

Dr. Christian Baumgartner is currently heading the Machine Learning for Medical Image Analysis Group which is part of the Cluster of Excellence: Machine Learning - New Perspectives for Science, at the University of Tübingen. Before joining the University of Tübingen, Christian was working in a senior research engineering role at PTC Vuforia, where he focused on research and development of machine learning technology for augmented reality applications. Prior to this, he was a Post-doc at the Biomedical Image Computing Group at ETH Zürich working with Prof. Ender Konukoglu, and before in the Biomedical Image Analysis Lab with Prof. Daniel Rueckert. Christian completed his PhD in 2016 under the joint supervision of Prof. Andy King and Prof. Daniel Rueckert at King’s College London in the School of Biomedical Engineering & Imaging Sciences. He obtained his Master’s degree in Biomedical Engineering and my Bachelor’s degree in Information Technology and Electrical Engineering from ETH Zürich.

More from the Same Authors