Deep Generative Models workshop @ MICCAI 2021


Deep generative models such as Generative Adversarial Network (GAN) and Variational Auto-Encoder (VAE) are currently receiving widespread attention from not only the computer vision and machine learning communities, but also in the MIC and CAI community. These models combine the advanced deep neural networks with classical density estimation (either explicit or implicit) for achieving state-of-the-art results. DGM4MICCAI workshop at MICCAI 2021 will be all about Deep Generative Models in Medical Image Computing and Computer Assisted Interventions.

Associated Challenge: AdaptOR

Workshop Proceedings available via Springer Link

DOI: 10.1007/978-3-030-88210-5

Submission Details

We seek contributions that include, but are not limited to:

We particularly welcome papers driven by the theme "MIC meets CAI". Interesting novel applications of deep generative models in MIC and CAI beyond these topics are also welcome.

Workshop proceedings are published as part of Springer Nature's Lecture Notes in Computer Science (LNCS) series. Manuscripts will be reviewed in double-blinded peer-review. Please prepare your workshop papers according to the MICCAI submission guidelines (LNCS template, 8 pages maximum).
Supplementary material: PDF documents or mp4 videos, 10 MB maximum file size.


Event Date
Paper Submission Deadline 25. June 2021 extended to 2. July
Review Release 18. July 2021
Rebuttal Due 21. July 2021
Final Decision 23. July 2021
DGM4MICCAI Workshop 1. October 2021


Start End Topic
2:00 PM2:10 PMWelcome
2:10 PM3:10 PMOral Session 1
3:10 PM3:55 PMKeynote: Stefanie Speidel
3:55 PM4:35 PMOral Session 2
4:35 PM5:20 PMKeynote: Andreas Maier
5:20 PM5:50 PMAnnouncing Winners of AdaptOR Challenge
5:50 PM6:00 PMClosing Remarks

All times are given in UTC time.

Download iCal Event

Keynote Speakers

Stefanie Speidel

Translational Surgical Oncology

NCT Dresden, Germany

Image-to-image translation for computer-assisted surgery – perspectives and challenges

Deep generative models have recently seen major success in various domains. While the methods evolve and further achievements in machine intelligence are expected, there exist numerous challenges in the medical domain that hamper a significant impact for healthcare.
The talk will broadly highlight current success stories of image-to-image translation for CAI. It will introduce challenges that are specific for CAI as well as possibilities to overcome them. Finally future directions will be discussed and strategies to bring the full potential to the patient.

Andreas Maier

Pattern Recognition Lab

FAU Nürnberg, Germany

Hybrid Magnet Resonance and X-ray imaging

In this talk, we will investigate the use of an interventional X-ray MRI System.
In particular, we focus on the joint use of fusing the advantages of fast X-ray imaging with the multitude of contrasts that are available in MRI.
On the one hand, we will explore methods using deep learning to compensate for the differences in geometry and on the other hand, generative methods that enable to map the one contrast onto the other.
Lastly, we will discuss hybrid imaging methods that enable us to unite the high resolution of X-ray images with the freedom of MRI contrasts efficiently.

Oral Sessions

Session 1

Paper IDTitleAuthors
1Frequency-Supervised MRI-to-CT Image SynthesisShi, Zenglin*; Mettes, Pascal; Zheng, Guoyan; Snoek, Cees
10Ultrasound Variational Style Transfer to Generate Images Beyond the Observed DomainHung, Alex*; Galeotti, John
133D-StyleGAN: A Style-Based Generative Adversarial Network for Generative Modeling of Three-Dimensional Medical ImagesHong, Sungmin*; Marinescu, Razvan V; Dalca, Adrian V; Bonkhoff, Anna K; Bretzner, Martin; Rost, Natalia; Golland, Polina
17Bridging the gap between paired and unpaired medical image translationPaavilainen, Pauliina*; Akram, Saad Ullah; Kannala, Juho
18Conditional generation of medical images via disentangled adversarial inferenceHavaei, Mohammad*; Mao, Ximeng; Wang, Yiping; Lao, Qicheng
19CT-SGAN: Computed Tomography Synthesis GANPesaranghader, Ahmad*; Wang, Yiping; Havaei, Mohammad

Session 2

Paper IDTitleAuthors
2Hierarchical Probabilistic Ultrasound Image Inpainting via Variational InferenceHung, Alex*; Sun, Zhiqing; Chen, Wanwen; Galeotti, John
11CaCL: class-aware codebook learning for weakly supervised segmentation on diffuse image patternsDeng, Ruining; Liu, Quan; Bao, Shunxing; Jha, Aadarsh; Chang, Catie; Millis, Bryan; Tyska, Matthew; Huo, Yuankai*
14BrainNetGAN: Data augmentation of brain connectivity using generative adversarial network for dementia classificationLi, Chao; Wei, Yiran; Chen, Xi*; Schönlieb, Carola-Bibiane
21Evaluating GANs in medical imagingTronchin, Lorenzo*; Sicilia, Rosa; Cordelli, Ermanno; Ramella, Sara; Soda, Paolo

AdaptOR Papers

Paper IDTitleAuthors
25Improved Heatmap-based Landmark DetectionYao, Huifeng*; Guo, Ziyu; Zhang , Yatao; Li, Xiaomeng
27Cross-domain Landmarks Detection in Mitral RegurgitationWang, Jiacheng; Wang, Haojie; Mu, Ruochen; Wang, Liansheng*