Maths meets Image
Hackathon on image reconstruction, segmentation and shape analysis
17 – 19 March 2022 at The Classroom, Berlin
Transforming the World
through Mathematics
Part of the Thematic Einstein Semester on Mathematics of Imaging in Real-World Challenges
Organizers (Young Academy Committee)
Felix Ambellan (ZIB)
Robert Beinert (TU Berlin)
Christoph Kolbitsch (PTB)
Kostas Papafitsoros (WIAS)
Christoph von Tycowicz (FU Berlin, ZIB)
Note regarding the COVID-19 pandemic
The Hackathon is a live event and requires 2G+. You will need to provide proof of vaccination or a valid medical certificate if you have recovered (don’t forget your ID). Otherwise, you need to show a valid negative test result for every day of the event. As we know regulations can change quickly depending on current infection rates. So please check this homepage regularly for any updates.
Mathematics is an integral part of imaging, ranging from image reconstruction, classification and segmentation to image interpretation. Often there is a gap between mathematical models and real-world data which hinders wide-spread application of exciting new research. In this 3-day-event we want to bridge this gap and bring together an interdisciplinary group of experts so that they can learn about imaging, gain first practical experience and address exciting new research questions. “Maths Meets Image” will take place from the 17th to the 19th of March, 2022 at the typical Berlin venue The Classroom. Participants should bring a laptop and lots of motivation, everything else will be provided by us.
What is a Hackathon?
Random people coming together and writing some more or less useful code. In our case we want to bring together young scientists from different research backgrounds working in image reconstruction, segmentation and classification. You will be working in teams of 5-6 people. We aim for the Hackathon to be the starting point for exciting future research collaborations.
This sounds like work. Why would I join?
Because it’s going to be fun, you’re going to get to know new people and you might even learn something which will be helpful for your own research. There is also going to be free food…
What is expected of me?
You are interested in imaging. You are not expected to be an expert in anything or even fully understand the project descriptions below. The Hackathon is about learning new things.
You have some experience in programming. No worries if you don’t know the difference between runtime and compile time polymorphism in C++ but some basic coding experience is needed.
You have a laptop. We will provide high performance computers, but you will need to bring your own laptop.
How do I join?
Please send us an email (young_academy@mathplus.de) and let us know which project you would like to join (see below). If you have a new idea for a project for the hackathon, feel free to share it with us. We are always curious about new ideas.
Below is a list of projects we aim to tackle during the Hackathon. If you have a new idea for a project for the Hackathon, feel free to share it with us. We are always curious about new ideas.
- Temporally dependent TV/TGV regularisation
Develop a TV/TGV regularised image reconstruction approach where the regularisation parameter is a function of space and time. The approach would be developed for 2D cardiac CINE imaging first using simulations (XCAT) and then real-data (radially acquired MR data). - Cardiac motion estimation using heart models
Utilise (dynamic) heart models to estimate the anatomical changes of the heart during the cardiac cycle. An approach would be developed to accurately estimate cardiac motion from 3D cardiac MR acquisitions. The final aim is to integrate the estimated cardiac motion in the MR image reconstruction algorithm to correct for it. - Estimation of 3D heart segmentation from sparse 2D MR images
Cardiac MRI is usually acquired using multiple 2D slices. The aim would be to investigate how a 3D heart segmentation can be obtained from 2D MR slices which can have gaps and/or are acquired in different orientations. - Morphological Scoring of Disease States
Obtain a classifier system that takes the non-Euclidean structure of shape spaces into account. The statistical shape model based approach [Ambellan et al. 2021] will serve as a starting point. Possible extensions include different metrics, advanced Riemannian mean-variance analysis, robust estimators, and intrinsic machine learning approaches. - Geometric deep learning for quantitative analysis of stone tool reduction sequences
Refitted core reduction nodules pose a promising source of information for understanding of variability in stone tool production — and thus cultural transmission — among prehistoric populations. This project targets the automatic detection of points of impact on lithic flakes respectively snapshots of the core sequence. To this end, state-of-the-art approaches form the field of geometric deep learning are to be adapted and experimentally evaluated. - Estimation of displacements and mechanical tissue properties from Magnetic Resonance Elastography data
MR Elastography is an imaging technique sensitive to tissue mechanics and tissue mechanical properties. Combined with mathematical models, it can be used to estimate biomarkers from low-resolution displacement data. This project targets the development of an automatic pipeline to obtain high-quality displacement images and estimate mechanical parameters applying smoothing and inverse estimators to the complex MRE signal.
The event will start on Thursday March 17, 9:00 am and conclude on Saturday March 19, 15:00 pm. The planned timetable is as follows:
Registration links will be sent to the participants via email. In the case you would like to participare and/or suggest a potential project, please get in contact with the organizers (young_academy@mathplus.de)
The Hackathon will take place at The Classroom, a unique Berlin-style venue in the center of the city. Drinks and food (vegetarian) will be provided during the whole event.
The address is Gerichtstrasse 23, Hof 4, Aufgang 4, 13347 Berlin.
The Hackathon is supported by:
- Collaborative Computational Project in Synergistic Reconstruction for Biomedical Imaging (https://www.ccpsynerbi.ac.uk)
- Collaborative Computational Project in Tomographic Imaging (https://www.ccpi.ac.uk)
- Prof. Dr.-Ing. Leonid Goubergrits (Head of Cardiovascular Modelling and Simulation Group, Charité – Universitätsmedizin Berlin and Professor at Einstein Center Digital Future)
- Prof. Dr.-Ing. Anja Hennemuth (Institute of Computer-assisted Cardiovascular Medicine, Charité – Universitätsmedizin Berlin)