Instructors: Prof. Dr. Nassir Navab, Dr. Shahrooz Faghih Roohi, Ashkan Khakzar, Azade Farshad, Roger Soberanis
- Registration must be done through TUM Matching Platform (please pay attention to the Deadlines)
- In order to increase your priority, please also apply via our own Registration system.
- The maximum number of participants: 20.
- The preliminary meeting slides can be found here: DLMASws21-22.pdf
- The preliminary meeting is scheduled for July 7th, 16:30 (Zoom link is visible on TUMonline).
- Due to the current pandemic, the seminar happens virtually via Zoom (the meeting link will be shared with participants via email).
- Deep Learning is growing tremendously in Computer Vision and Medical Imaging as well. Highly impacted journals in the medical imaging community, i.e. IEEE Transaction on Medical Imaging, published recently their special edition on Deep Learning . The Seminar will propose a list of recent scientific articles related to the main current research topics in deep learning for Medical Applications together with some interesting papers from other communities (CVPR, NeurIPS, ICCV, ICLR, ICML, ...).
In this Master Seminar (Hauptseminar), students select one scientific article from the list provided by course organizers. The students read the paper, and must accomplish the following:
- Presentation: The selected paper is presented to the other participants (20 minutes presentation 10 minutes questions). You can use the CAMP templates for PowerPoint camp-tum-jhu-slides.zip, or Latex: CAMP-latex-template.
- Blog Post: A blog post of 1000-1200 words excluding references should be submitted before the deadline. The blog post must include all references used and must be written completely in your own words. Copy and paste will not be tolerated
- Attendance: Participants have to participate actively in all seminar sessions. Each presentation is followed by a discussion and everyone is encouraged to actively participate.
Submission Deadline: You have to submit the blog post one week before your presentation session! You should also submit the presentation one day right after your presentation session.
You may look at the blog posts and presentations from the previous semester (using the tree in the upper left corner).
List of Topics and Material
Papers in this course are selected from the following venues/publications:
CVPR: Conference on Computer Vision and Pattern Recognition
ICLR: International Conference on Learning Representations
NeurIPS: Neural Information Processing Systems
TPAMI: IEEE Transactions on Pattern Analysis and Machine Intelligence
TMI: IEEE Transaction on Medical Imaging
JBHI: IEEE Journal of Biomedical and Health Informatics
MedIA: Medical Image Analysis (Elsevier)
MICCAI: Medical Image Computing and Computer-Assisted Intervention
BMVC: British Machine Vision Conference
MIDL: Medical Imaging with Deep Learning
List of papers
|Medical Image segmentation / Registration/ Report generation||1||Semi-supervised Medical Image Segmentation through Dual-task Consistency||AAAI 2021||Shahrooz||https://ojs.aaai.org/index.php/AAAI/article/view/17066|
|2||Exploring and Distilling Posterior and Prior Knowledge for Radiology Report Generation||CVPR 2021||Matthias|
|3||Anatomy-Constrained Contrastive Learning for Synthetic Segmentation Without Ground-Truth||MICCAI 2021||Mehdi||https://link.springer.com/chapter/10.1007%2F978-3-030-87193-2_5|
|4||Comparison of thyroid segmentation techniques for 3D ultrasound||Medical Imaging 2017: Image Processing||Chrissi, Lennart, Matthias||https://www.var.ovgu.de/pub/2017_Wunderling_SPIE_comparison-thyroid-segmentation-SPIE-submission.pdf|
|Unsupervised/Semi-supervised/Self-supervised learning||5||Unpaired Training of Deep Learning tMRA for Flexible Spatio-Temporal Resolution||TMI 2021||Shahrooz||https://ieeexplore.ieee.org/abstract/document/9195022|
|6||IN DEFENSE OF PSEUDO-LABELING: AN UNCERTAINTY-AWARE PSEUDO-LABEL SELECTION FRAMEWORK FOR SEMI-SUPERVISED LEARNING||ICLR 2021||Tariq||https://openreview.net/pdf?id=-ODN6SbiUU|
|7||Local plasticity rules can learn deep representations using self-supervised contrastive predictions||NeurIPS 2021||Ashkan||https://arxiv.org/abs/2010.08262|
|8||Prototypical Contrastive Learning of Unsupervised Representations||ICLR 2021||Tariq||https://arxiv.org/pdf/2101.06329v3.pdf|
|9||Improving Contrastive Learning by Visualizing Feature Transformation||ICCV 2021||Yousef|
|10||How Well Do Self-Supervised Models Transfer?||CVPR 2021||Yousef|
|11||Dense Contrastive Learning for Self-Supervised Visual Pre-Training||CVPR 2021||Yousef|
|12||Object-aware Contrastive Learning for Debiased Scene Representation||arXive 2021 (Google)||Yousef||https://arxiv.org/pdf/2108.00049v1.pdf|
|Data-Efficient DL, Representation Learning||13||Public Covid-19 X-ray datasets and their impact on model bias – A systematic review of a significant problem||MedIA 2021||Ashkan||https://www.medrxiv.org/content/10.1101/2021.02.15.21251775v1.full|
|14||Shallow Bayesian Meta Learning for Real-World Few-Shot Recognition||ICCV 2021||Azade||https://arxiv.org/pdf/2101.02833.pdf|
|15||Exploring Simple Siamese Representation Learning||CVPR 2021||Azade|
|16||Conditional Deformable Image Registration with Convolutional Neural Network||MICCAI 2021||Anees|
|Reliable AI (Interpretability, Uncertainty, robustness)||17||Right for the Right Concept: Revising Neuro-Symbolic Concepts by Interacting with their Explanations||CVPR 2021||Ashkan||https://arxiv.org/abs/2011.12854|
|18||Propagating Uncertainty Across Cascaded Medical Imaging Tasks For Improved Deep Learning Inference||TMI 2021||Shahrooz||https://ieeexplore.ieee.org/abstract/document/9541203|
|19||How Well do Feature Visualizations Support Causal Understanding of CNN Activations?||NeurIPS 2021||Ashkan|
|20||Generative Classifiers as a Basis for Trustworthy Computer Vision||CVPR 2021||Heiko|
|Transformers||21||DEFORMABLE DETR: DEFORMABLE TRANSFORMERS FOR END-TO-END OBJECT DETECTION||ICLR2021||Farid||https://arxiv.org/pdf/2010.04159.pdf|
|22||TransPath: Transformer-Based Self-supervised Learning for Histopathological Image Classification||MICCAI 2021||Anees|
|23||Rethinking Semantic Segmentation from a Sequence-to-Sequence Perspective with Transformers||CVPR 2021||Yousef|
|24||Transformer Interpretability Beyond Attention Visualization||CVPR 2021||Yousef|
|Graphs||25||DualGraph: A graph-based method for reasoning about label noise||CVPR 2021||Mahsa|
|26||ON GRAPH NEURAL NETWORKS VERSUS GRAPH AUGMENTED MLPS||ICLR 2021||Mahsa||https://openreview.net/pdf?id=tiqI7w64JG2|
|27||Heterogeneous Graph Transformer||ACM DL||Anees||https://arxiv.org/abs/2003.01332|
|Other||28||pi-GAN: Periodic Implicit Generative Adversarial Networks for 3D-Aware Image Synthesis||CVPR 2021||Farid|
|29||Neural Deformation Graphs for Globally-consistent Non-rigid Reconstruction||CVPR 2021||Farid|
|30||Deep learning on ultrasound images of thyroid nodules||Biocybernetics and Biomedical Engineering||Chrissi, Lennart, Matthias||https://www.sciencedirect.com/science/article/abs/pii/S0208521621000152|
|31||Bayesflow: Learning Complex Stochastic Models with Invertible Neural Networks||IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS 2020||Heiko||https://ieeexplore.ieee.org/document/9298920|
|32||Model-Contrastive Federated Learning||CVPR 2021||Yousef|
|33||Autonomic Robotic Ultrasound Imaging System Based on Reinforcement Learning||IEEE Transactions on Biomedical Engineering||Maria||https://ieeexplore.ieee.org/xpl/RecentIssue.jsp?punumber=10|
Literature and Helpful Links
A lot of scientific publications can be found online.
The following list may help you to find some further information on your particular topic:
- Microsoft Academic Search
- Google Scholar
- Collection of Computer Science Bibliographies
Libraries (online and offline):
- http://rzblx1.uni-regensburg.de/ezeit/ (Elektronische Zeitschriften Bibliothek)
- Verbundkatalog des Bibliotheksverbundes Bayern (BVB)
- Computer ORG
- http://www.ub.tum.de/ (TUM Library)
- Various proceedings of conferences in our AR-Lab, 03.13.036 (These proceedings are not for lending!)
Some further hints for working with references:
- JabRef is a Java program for comfortable working with Bibtex literature databases. Handy feature: if you know the PubMed ID for an article, JabRef can import data from there (via "Web Search/Medline").
- Mendeley is a cross-platform program for organising your references.
If you find useful resources that are not already listed here, please tell us, so we can add them for others. Thanks.