Recent new papers on graph-based learning and point cloud registration published at MICCAI 2023, ICCV 2023, NeurIPS 2023 and Medical Image Analysis

A. Bigalke, .., MP Heinrich:
Anatomy-guided domain adaptation for 3D in-bed human pose estimation (pdf Elsevier)
Medical Image Analysis2023

MP Heinrich, A. Bigalke, .. L. Hansen:
Chasing clouds: Differentiable volumetric rasterisation of point clouds .. for .. 3D registration GitHub Code
International Conference on Computer Vision ICCV oral 2023

A. Bigalke, L. Hansen, TCW. Mok, MP. Heinrich:
Unsupervised 3D registration through optimization-guided cyclical self-trainingGitHub Code
Medical Image Computing and Computer-Assisted Interventions MICCAI 2023

F. Falta, .., MP Heinrich:
Lung250M-4B: A Combined 3D Dataset for CT- and Point Cloud-Based Intra-Patient Lung Registration GitHub Page
Neural Information Processing Systems NeurIPS 2023

A. Hering, L Hansen, .. MP Heinrich:
Learn2Reg: comprehensive multi-task medical image registration challenge, dataset and evaluation in the era of deep learning pdf open access
IEEE Trans. Medical Imaging 2023

M. Blendowski, L. Hansen, MP Heinrich:
Weakly-supervised learning of multi-modal features for regularised iterative descent in 3D .. registration(pdf author)
Medical Image Analysis 2021

L Hansen, MP Heinrich:
GraphRegNet: Deep Graph Regularisation Networks on Sparse Keypoints (pdf author)
IEEE Trans. Medical Imaging 2021

B. Kainz, MP Heinrich, ..:
Non-invasive diagnosis of deep vein thrombosis from ultrasound imaging with machine learning(pdf open access)
Nature pj Digital Medicine 2022

MP Heinrich, L Hansen: Highly accurate and memory efficient unsupervised learning-based discrete CT registration using 2.5D displacement search please see paper for details. And our GitHub repository for open-source code

Learn2Reg registration challenge accepted as workshop at MICCAI 2020, 2021, 2022, 2023 please see website for details.

Recent accepted publications at MICCAI 2019, MIDL 2019:

Code for discrete deep learning registration pdd-net
MP Heinrich: Closing the Gap between Deep and Conventional Image Registration using Probabilistic Dense Displacement Networks MICCAI 2019

Code for self-supervised 3D fearures
M Blendowski, H Nickisch, MP Heinrich: How to learn from unlabeled volume data: Self-Supervised 3D Context Feature Learning MICCAI 2019

M Blendowksi, MP Heinrich: Learning interpretable multi-modal features for alignment with supervised iterative descent MIDL 2019 oral presentation

There are a number of fully-funded (80-100%) new positions as research assistant (PhD student or PostDoc) available in my group please see opportunities

Recent accepted publications at MICCAI 2018, MIDL 2018, IEEE TBME and IJCARS journals:

Free Access to our new Medical Image Analysis Paper
OBELISK-Net: Fewer Layers to Solve 3D Multi-Organ Segmentation with Sparse Deformable Convolutions

MP Heinrich, O Oktay, N Bouteldja: OBELISK – One Kernel to Solve Nearly Everything: Unified 3D Binary Convolutions for Image Analysis MIDL 2018 paper (oral presentation video)
(for more information and code, see GitHub)
Winner of Best Paper Award (see MIDL Website, and invited Medical Image Analysis paper)

R Tanno, A Makropoulos, S Arslan, O Oktay, S Mischkewitz, F Al-Noor1, J Oppenheimer, R Mandegaran, B Kainz, MP Heinrich AutoDVT: Joint Real-time Classification for Vein Compressibility Analysis in Deep Vein Thrombosis Ultrasound Diagnostics MICCAI 2018

IY Ha, M Wilms, H Handels, MP Heinrich Model-based Sparse-to-dense Image Registration for Realtime Respiratory Motion Estimation in Image-guided Interventions IEEE Trans Biomedical Imaging
(for more information and code, see GitHub)

MP Heinrich, M Blendowski, O Oktay TernaryNet: Faster Deep Model Inference without GPUs for Medical 3D Segmentation using Sparse and Binary Convolutions Int J Computer Assisted Radiology and Surgery (MICCAI 2018 Special Issue)
(for more information and code, see GitHub)

Our paper on efficient deep CNNs with binary sparse convolutions has been accepted for MICCAI 2017 in Quebec:
MP Heinrich, O Oktay: BRIEFnet: Deep Pancreas Segmentation using Binary Sparse Convolutions (pdf)
(for more information and code, see

A new state-of-the-art in deformable image registration (first rank in EMPIRE10) has been reached in our new IEEE Transactions on Medical Imaging paper:
J Rühaak, T Polzin, S Heldmann, IJA Simpson, H Handels, J Modersitzki, MP Heinrich:
Estimation of Large Motion in Lung CT by Integrating Regularized Keypoint Correspondences into Dense Deformable Registration
fissure segmentation data for COPD scans is available at research
for a short overview see ieee-tmi-org

We presented three papers at MICCAI 2016 in Athens (including):
MP Heinrich, M Blendowski: Multi-Organ Segmentation using Vantage Point Forests and Binary Context Features (pdf)
(for more information and code, see research)

Special Issue on Graphical Models in Medical Image Analysis:
MP Heinrich, IJA Simpson, BW Papiez, M Brady, JA Schnabel: Deformable image registration by combining uncertainty estimates from supervoxel belief propagation(link to Elsevier) Code is coming soon

Medical Image Analysis 20th Anniversary Issue:
JA Schnabel, MP Heinrich, BW Papiez, M Brady: Advances and Challenges in Deformable Image Registration: From Image Fusion to Complex Motion Modelling (link to Elsevier)

dphil graduationWhen graduating from Oxford, you get to wear a neat gown.


My discrete registration approach deeds (see software) has won the first place in a comprehensive abdominal registration comparison against NiftyReg, ANTS SyN, IRTK and FSL. The paper describing the results is entitled "Evaluation of Six Registration Methods for the Human Abdomen on Clinically Acquired CT" (by Z Xu, C Lee, MP Heinrich, M Modat, D Rueckert, Ourselin S, R Abramson, and B Landman) at IEEE TBME. Details on the parameter settings for deeds can be found in the following short paper (pdf) and presentation (talk).

I have won the British Machine Vision Association and Society for Pattern Recognition (BMVA) Sullivan Thesis Prize 2014 for the best UK PhD thesis in Computer Vision. More details can be found at:

My mini-lecture for the MICCAI Educational Challenge on Random Walk Segmentation (2nd Prize, Jury Vote) can be found here MEC 2015 or directly here QuickTime Movie (Resolution: 1152x720, AAC, H.264, duration 8 min 22 secs, size 59 MByte)

About me

In August 2013, I finished my D.Phil. (Ph.D.) at the University of Oxford. I used to work at the Institute of Biomedical Engineering together with Julia Schnabel, Mark Jenkinson and Sir Michael Brady. My thesis was entitled: "Deformable Lung Registration for Pulmonary Image Analysis of MRI and CT scans".

In March 2019, I have been promoted to the position of associate professor at the Institute of Medical Informatics (IMI), at the University of Luebeck. In my group of 8 researcher, we work on medical deep learning, e.g. with application to image registration and motion compensation, deep learning based segmentation, probabilistic methods, self-supervised learning and many other useful things.

bodtener uferLuebeck is just 20 min. away from the Baltic Sea.

My main research focus lies in the development of deep learning and deformable image registration. During the last years, I mainly worked on CT and MRI registration and segmentation, but I am also very interested in multi-modal image learning. In my work, I always aim to develop methods that are computational very efficient and also widely applicable. Previously, I developed novel descriptor-based image representations, which can improve the robustness and accuracy of a large variety of medical image registration tasks. Further, I developed discrete optimisation strategies, which can greatly reduce the time needed for deformable image registration to less than a minute, making the software useful for time-sensitive applications and large datasets.

If you would like to use any of the code I developed for my previous publications, are interested in collaborating on a related medical image analysis project or are looking for a supervisor for a student project (B.Sc./M.Sc.), please don't hesitate to contact me.

Contact Details:

Dr. Mattias P. Heinrich,
Insitute of Medical Informatics, Universität Lübeck
Ratzeburger Allee 160, 23538 Lübeck, Germany
Tel.: +49-451-3101-5602
heinrich (at) imi (dot) uni-luebeck (dot) de