Bjoern Menze (TUM), Markus Kowarschik.
Type: Master Seminar Module IN2107
SWS: 2+0
ECTS: 5 Credits
Location: Munich School of Bioengineering (formerly: IMETUM) Seminar Room
Time: Wednesday 1300-1500
Course Language: English
Many clinical imaging sequences include image modalities that provide unique insights into the human body: describing blood flow pattern or water diffusion, visualizing micro-structural tissue properties, or quantifying metabolic processes that have been found to be relevant for specific diseases. To access and explore the rich biological information of such multimodal CT, PET, and MR image data sets, different approaches from biophysical modeling and computational physiology are used.
The topic of this seminar is to survey such computational models by reading, discussing and understanding key publications in this field, for example, on:
Emphasis is put on improving presentation style. All presenters will present their talk twice during the course of the semester, and after the first time the group discussion will focus on presentation style.
Each presentation is followed by a discussion and everyone is encouraged to actively participate. The handout must include all references used and should be your own review of the paper (don't copy & paste). Both handout and presentation will have to be in English. It can follow the structure of the presentation: Intro (incl. statement of the problem, prior work, contribution of the present work), Methods (previously existing and improvements), Experiments (data and results), Conclusion / Summary.
The seminar consists of an introductory meeting and seminar sessions. Each of the sessions will consist of several student presentations. All seminar participants are expected to be present at every session and to actively participate in the discussion.
The papers that we are going to discuss in the seminar belong to four different research areas:
Area | Description | Presenter | Tutor ![]() |
---|---|---|---|
F Ozdemir et al. Learn th enew, keep the old: extending pretrained models with new anatomy and images. Miccai 2018 | Jorge | Markus | |
Shape & registration | B Becker et al. Deep multi-structural shape analysis: application to neuroanatomy. Miccai 2018 | Zlatija | Markus |
V Kong et al. Invasive cancer detection utilizing compressed convolutional neural networks and transfer learning. Miccai 2018 | Chris | Bjoern | |
S Singla et al. Subject2vec: generative-discriminative approach from a bag of image patches to a vector. Miccai 2018 | Rameshwara | Bjoern | |
Uncertainty, sparse models | Nair et al. Exploring uncertainty measures in deep networks for multile sclerosis lesion detection and segmentations. Miccai 2018 | ||
F Bragman et al. Uncertainty in multitask learning: joint representations for probabilistic MR only radiotherapy planning. Miccai 2018 | |||
M Paschali et al. Generalizability vs. robustness: adversarial examples for medical iamging. Miccai 2018 | |||
A Rost et al. Inherent brain segmentation quality control from fully ConvNet? Monte Carlo sampling. Miccai 2018 | |||
M Kozinski et al. Learning to segment 3D linear structures using only 2D annotations. Miccai 2018 | |||
Networks, curvilinear structures | M Meyer et al. A pixel-wise distance regression approach for joint retinal optincal disc and fovea detection. Miccai 2018 | ||
A Galdran et al. A no-reference retinal vessel tree segmentation quality metric. Miccai 2018 | |||
S Moriconi et al. Elastic registration of geodesic vascular graphs. Miccai 2018 | |||
Y Cid et al. From local to global: a holistic lung graph model. Miccai 2018 | |||
Z Zgau et al. Pulmonary vessel tree matching for quantifying changes in vascular morphology. Miccai 2018 | |||
R Selvan et al. Mean field network based graph refinement with application to airway tree extraction. Miccai 2018 | |||
Multi-organ segmentation | Y Wang et al. Training multi-organ segmentation networks with sample selection by relaxed upper confident bound. Miccai 2018 | ||
M Vakalopoulos et al. AtlasNet?: multi-altas non-linear deep networks for medical image segmentations. Miccai 2018 | |||
K Won et al. 3d segmentation with exponential logarithmic loss for highly unbalanced object sizes. Miccai 2018 | |||
Signal models & physiology | C Scott et al. Short acquisition time PET/MR pharmacokinetic modeling using CNNs. Miccai 2018 | ||
M Molero et al. Deep learning with synthetic diffusion MRI data for free-water elimination in glioblastoma cases. Miccai 2018 | |||
J Schlemper et al. Cardiac MR segmentation from undersampled k-space using deep latent representation learning. Miccai 2018 | |||
S Bulmberg et al. Deeper image quality transfer: retraining low-memory neural etworks for 3D images. Miccai 2018 | |||
S Bayer et al. Intraoperative brain shift compensation using a hybrid mixture model. Miccai 2018 | |||
Text & images | J Dockes et al. text to brain: predicting the spatial distribution of neuroimaging observations from text reports. Miccai 2018 | ||
A Dalco et al. Unsupervised learning for fast probabilistic diffeomarphic registration. Miccai 2018 |
The seminar consists of an introductory meeting and seminar sessions. Each of the sessions will consist of several student presentations. All seminar participants are expected to be present at every session and to actively participate in the discussion.
A lot of scientific publications can be found online.
The following list may help you to find some further information on your particular topic:
Some publishers:
Libraries (online and offline):
Some further hints for working with references:
If you find useful resources that are not already listed here, please tell us, so we can add them for others. Thanks.
TeachingForm | |
---|---|
Title: | Image based biomedical modeling seminar |
Professor: | Bjoern Menze |
Tutors: | |
Type: | Hauptseminar |
Information: | Hauptseminar, SWS: 2, ECTS: 5 |
Term: | 2018WiSe |
Abstract: |