Study Groups

New to MIDL 2021 and based on its success in recent IPMI conferences, we are introducing Study Groups: The mentor for a study group will meet for 45 minutes with a number of PhD students and jointly decide what questions might be of most interest for each paper in a session. This stimulates the live discussion and encourages younger MIDL members to actively participate. In addition, mentors can pass on some of your knowledge about the state-of-the-art in the field, the review process at MIDL, career advice and future challenges for research.

For attendees, the help of the mentor will support them in coming up with intriguing questions to ask at these group discussions. For many early-career scientists, it may be overwhelming to speak up yourself during a live session - being part of a study group makes it easier and also helps them meet new people in the field.

While each group may flexibly decide upon an exact meeting time and space, we have allocated two slots per day in our conference schedule and special "study group" rooms in Gather.Town. These are 11:45 to 12:30 (UTC+2) for Europe/Africa/Asia, .. and 17:45 to 18:30 (ET, UTC-4) for Americas. See Program

Participants of study groups will receive an individual email about their one assigned session.

Please note that the session chairs of the corresponding live discussion below are only included for completeness and not present during the study group.

Study Groups Tuesday 20.00-20.45 (UTC+2)

Study Group 1-A

• Mentor: Francesco Caliva (4+1)
• Session Chairs (A4-12): Francesco Caliva, Christian Desrosies

This study group discusses the following papers:

• A6: Christian Schiffer - Learning to predict cutting angles from histological human brain sections - (short)
• A7: Markus Philipp - Localizing neurosurgical instruments across domains and in the wild - (long)
• A8: Qian He - Weakly Supervised Volumetric Segmentation via Self-taught Shape Denoising Model - (full)
• A9: Andreanne Lemay - Benefits of Linear Conditioning for Segmentation using Metadata - (full)
• A10: Harsh Maheshwari - Distill DSM: Computationally efficient method for segmentation of medical imaging volumes - (full)

Study Group 1-B:

• Mentor: Vincent Andrearczyk (3+3)
• Session Chairs: Sandy Engelhardt, Lena Maier-Hein

This study group discusses the following papers

• C1: Andreas M Kist - Efficient biomedical image segmentation on Edge TPUs - (short)
• C2: Alessandra Lumini - Deep ensembles based on Stochastic Activation Selection for Polyp Segmentation - (short)
• C3: Javier Morlana - Self-supervised Visual Place Recognition for Colonoscopy Sequences - (short)
• C5: Bokai Zhang - SWNet: Surgical Workflow Recognition with Deep Convolutional Network- (full)
• C6: Daniel Neimark - “Train one, Classify one, Teach one” - Cross-surgery transfer learning for surgical step recognition - (full)
• C7: Aishik Konwer - Predicting COVID-19 Lung Infiltrate Progression on Chest Radiographs Using Spatio-temporal LSTM based Encoder-Decoder Network - (full)

Study Group 1-C:

• Mentor: Hoel Kervadec (3+2)
• Session Chairs (K Friday: Chairs: Hoel Kervadec, Max-Heinrich Laves)

This study group discusses the following papers

• A11: Kimberly Amador - Stroke Lesion Outcome Prediction Based on 4D CT Perfusion Data Using Temporal Convolutional Networks - (full)
• A12: Philipp Gruening - Direct Inference of Cell Positions using Lens-Free Microscopy and Deep Learning - (full)
• K4: Soumick Chatterjee - DS6, Deformation-aware Semi-supervised Learning: Application to Small Vessel Segmentation with Noisy Training Data - (short)
• K5: Jan Mikolaj Kaminski - Deep ensemble model for segmenting microscopy images in the presence of limited labeled data - (short)
• K6: Md Asadullah Turja - Learning the Latent Heat Diffusion Process through Structural Brain Network from Longitudinal β-Amyloid Data - (full)

Study Group 1-D:

• Mentor: Jannis Hagenah (3+3)
• Session Chairs D1-3: Jannis Hagenah, Caroline Petitjean; E Chairs: Alessa Hering, Hervé Lombaert

This study group discusses the following papers

• D1: Christian Abbet - Self-Rule to Adapt: Learning Generalized Features from Sparsely-Labeled Data Using Unsupervised Domain Adaptation for Colorectal Cancer Tissue Phenotyping
• D2: Steffen Czolbe - Semantic similarity metrics for learned image registration
• D3: Chao Feng - Nuc2Vec: Learning Representations of Nuclei in Histopathology Images with Contrastive Loss
• E10: Shibo Xing - Cycle Consistent Embedding of 3D Brains with Auto-Encoding Generative Adversarial Networks - (short)
• E11: Leihao Wei - Efficient and Accurate Spatial-Temporal Denoising Network for Low-dose CT Scans - (short)
• E12: Joon-Ho Son - Synthesis of Diabetic Retina Fundus Images Using Semantic Label Generation - (short)

Study Group 1-E:

• Mentor: Alessa Hering (1+5)
• Session Chairs: Alessa Hering, Hervé Lombaert

This study group discusses the following papers

• E4: Mattias P Heinrich - Rethinking the Design of Learning based Inter-Patient Registration using Deformable Supervoxels - (short)
• E5: henrik skibbe - Semi-supervised Image-to-Image translation for robust image registration - (short)
• E6: Junyu Chen - ViT-V-Net: Vision Transformer for Unsupervised Volumetric Medical Image Registration - (short)
• E7: Hanna Siebert - Learning a Metric without Supervision: Multimodal Registration using Synthetic Cycle Discrepancy - (short)
• E8: Huaqi Qiu - Learning Diffeomorphic and Modality-invariant Registration using B-splines - (full)
• E9: Joshua Russell Astley - A hybrid model- and deep learning-based framework for functional lung image synthesis from non-contrast multi-inflation CT - (short)

Study Groups Wednesday 11:45-12:30 (UTC+2)

Study Group 2-A:

• Mentor: Jelmer Wolterink (3+2)
• Session Chairs: Minjeong Kim, Jelmer Wolterink

This study group discusses the following papers

• A1: Alessa Hering - Whole-Body Soft-Tissue Lesion Tracking and Segmentation in Longitudinal CT Imaging Studies
• A2: Manan Lalit - Embedding-based Instance Segmentation in Microscopy
• A3: Hoel Kervadec - Beyond pixel-wise supervision: semantic segmentation with higher-order shape descriptors
• A4: Annika Reinke - Common limitations of performance metrics in biomedical image analysis - (short)
• A5: Hoang Canh Nguyen - VinDr-RibCXR: A Benchmark Dataset for Automatic Segmentation and Labeling of Individual Ribs on Chest X-rays - (short)

Study Group 2-B:

• Mentor: Francesco Ciompi (1+5)
• Session Chairs: B Mitko Veta, Jianhua Yao; Chairs D4-11: Tal Arbel, Hans Meine

This study group discusses the following papers

• B1: Jeroen Vermazeren - $\mu$PEN: Multi-class PseudoEdgeNet for PD-L1 assessment - (short)
• B2: Petr Kuritcyn - Comparison of CNN models on a multi-scanner database in colon cancer histology - (short)
• B3: Loris Nanni - Exploiting Adam-like Optimization Algorithms to Improve the Performance of Convolutional Neural Networks - (short)
• D8: Bram de Wilde - Cine-MRI detection of abdominal adhesions with spatio-temporal deep learning - (short)
• D9: Jonathan Rubin - Efficient Video-Based Deep Learning for Ultrasound Guided Needle Insertion - (short)
• D10: Dominik Mairhöfer - An AI-based Framework for Diagnostic Quality Assessment of Ankle Radiographs - (full)

Study Group 2-C:

• Mentor: Geert Litjens (1+5)
• Session Chairs: B Mitko Veta, Jianhua Yao

This study group discusses the following papers

• B4: Shana Beniamin - Gated CNNs for Nuclei Segmentation in H&E Breast Images - (short)
• B5: Kristina Lynn Kupferschmidt - Strength in Diversity: Understanding the impacts of diverse training sets in self-supervised pre-training for histology images - (short)
• B6: Nick Pawlowski - Abnormality Detection in Histopathology via Density Estimation with Normalising Flows - (short)
• B7: Yinan Zhang - Learning to Represent Whole Slide Images by Selecting Cell Graphs of Patches - (short)
• B8: Gijs Smit - Quality control of whole-slide images through multi-class semantic segmentation of artifacts - (short)
• B9: Yash Sharma - Cluster-to-Conquer: A Framework for End-to-End Multi-Instance Learning for Whole Slide Image Classification - (full)

Study Group 2-D

• Mentor: Colin Jacobs: (3+2)
• Session Chairs L (Friday): Christian Baumgartner, Colin Jacobs, Chairs D4-11: Tal Arbel, Hans Meine

This study group discusses the following papers

• D6: Francesco Caliva - Virtual Bone Shape Aging - (short)
• D7: Ansh Kapil - Breast cancer patient stratification using domain adaptation based lymphocyte detection in HER2 stained tissue sections - (short)
• D11: Juan Carlos Prieto - Image Sequence Generation and Analysis via GRU and Attention for Trachomatous Trichiasis Classification - (full)
• L5: Xiaobin Hu - Feedback Graph Attention Convolutional Network for MR Images Enhancement by Exploring Self-Similarity Features - (full)
• L6: Pauline Mouches - Unifying Brain Age Prediction and Age-Conditioned Template Generation with a Deterministic Autoencoder - (full)

Study Group 2-E

• Mentor: Anirban Mukhopadhyay (3+3)
• Session Chairs: Sandy Engelhardt, Lena Maier-Hein; Chairs D4-11: Tal Arbel, Hans Meine

This study group discusses the following papers

• C4: Raghavendra Selvan - Carbon footprint driven deep learning model selection for medical imaging - (short)
• C8: Andreas M Kist - Feature-based image registration in structured light endoscopy - (full)
• C9: Oleh Dzyubachyk - Intensity Correction and Standardization for Electron Microscopy Data - (full)
• D4: Tianyu Zhang - Predicting molecular subtypes of breast cancer using multimodal deep learning and incorporation of the attention mechanism - (short)
• D5: Yuchen Yang - Double adversarial domain adaptation for whole-slide-imageclassification - (short)
• D12: Hari Sowrirajan - MoCo Pretraining Improves Representation and Transferability of Chest X-ray Models - (full)

Study Groups Wednesday 23:45-0:30 (UTC+2)

Study Group 3-A

• Mentor: Raghav Mehtaa (2+4)
• Session Chairs H:: Raghav Mehta, Clarisa Sanchez

This study group discusses the following papers

• H6: Lucas de Vries - Transformers for Ischemic Stroke Infarct Core Segmentation from Spatio-temporal CT Perfusion Scans - (short)
• H7: Jueqi Wang - Multichannel input pixelwise regression 3D U-Nets for medical image estimation with 3 applications in brain MRI - (short)
• H8: Mickael Tardy - Morphology-based losses for weakly supervised segmentation of mammograms - (short)
• H9: Laura Hellwege - Partial Convolution Network for Metal Artifact Reduction in CT Preprocessing: Preliminary Results - (short)
• H10: Rosana EL Jurdi - A Surprisingly Effective Perimeter-based Loss for Medical Image Segmentation - (full)
• ~~H11: Hang Zhang - Memory U-Net: Memorizing Where to Vote for Lesion Instance Segmentation - (full)~~

Study Group 3-B

• Mentor: Christian Baumgartner (3+1)
• Session Chairs L (Friday): Christian Baumgartner, Colin Jacobs

This study group discusses the following papers

• L4: Philippe Weitz - Prediction of Ki67 scores from H&E stained breast cancer sections using convolutional neural networks - (short)
• L7: Chanh Nguyen - GOAL: Gist-set Online Active Learning for Efficient Chest X-ray Image Annotation - (full)
• L8: Hassan Muhammad - EPIC-Survival: End-to-end Part Inferred Clustering for Survival Analysis, with Prognostic Stratification Boosting - (full)
• L9: Soham Uday Gadgil - CheXseg: Combining Expert Annotations with DNN-generated Saliency Maps for X-ray Segmentation - (full)

Study Groups Thursday 11:45-12:30 (UTC+2)

Study Group 4-A

• Mentor: Christian Ledig (3+3)

• Session Chairs G: Angelica Aviles-Rivero, Antoine Théberge This study group discusses the following papers

• G4: Hansang Lee - Test-Time Mixup Augmentation for Uncertainty Estimation in Skin Lesion Diagnosis - (short)

• G5: Ke Yan - Interpretable Medical Image Classification with Self-Supervised Anatomical Embedding and Prior Knowledge - (short)
• G6: Polina Druzhinina - 50 shades of overfitting: towards MRI-based neurologicalmodels interpretation - (short)
• G7: Kangning Liu - Weakly-supervised High-resolution Segmentation of Mammography Images for Breast Cancer Diagnosis - (full)
• G8: Matthäus Heer - The OOD Blind Spot of Unsupervised Anomaly Detection - (full)
• G9: Hongrun Zhang - A regularization term for slide correlation reduction in whole slide image analysis with deep learning - (full)

Study Group 4-B

• Mentor: Hugo Kuijf: (3+0)
• Session Chairs: Angelica Aviles-Rivero, Antoine Théberge

This study group discusses the following papers

• H1: Jinwei Zhang - Hybrid optimization between iterative and network fine-tuning reconstructions for fast quantitative susceptibility mapping
• H2: Malte Tölle - A Mean-Field Variational Inference Approach to Deep Image Prior for Inverse Problems in Medical Imaging
• H3: Carolin Pirkl - Residual learning for 3D motion corrected quantitative MRI: Robust clinical T1, T2 and proton density mapping

Study Group 4-C

• Mentor: Ivana Isgum (3+3)
• Session Chairs E: Ivana Isgum, Carole Sudre Chairs F: Nicha Dvornek, Bram van Ginneken

This study group discusses the following papers

• E1: David A. Wood - Automated triaging of head MRI examinations using convolutional neural networks
• E2: Antoine Olivier - Balanced sampling for an object detection problem - application to fetal anatomies detection
• E3: Walter Hugo Lopez Pinaya - Unsupervised Brain Anomaly Detection and Segmentation with Transformers
• F1: Alexander M. Zolotarev - Ex-vivo - to - In-vivo Learning in Cardiology - (short)
• F2: Paul Weiser - Reconstruction and coil combination of undersampled concentric-ring MRSI data using a Graph U-Net - (short)
• F3: Soumick Chatterjee - ReconResNet: Regularised Residual Learning for MR Image Reconstruction of Undersampled Cartesian and Radial Data - (short)

Study Group 4-D

• Mentor: Bram van Ginneken (1+5)
• Session Chairs F: Nicha Dvornek, Bram van Ginneken

This study group discusses the following papers

• F4: Mikhail Bortnikov - 3D Scout Scans Using Projection Domain Denoising - (short)
• F5: Emanoel Ribeiro Sabidussi - Recurrent Inference Machines as Inverse Problem Solvers for MR Relaxometry - (short)
• F6: Farah Shamout - An artificial intelligence system for predicting the deterioration of COVID-19 patients in the emergency department - (short)
• F7: Ahmad Wisnu Mulyadi - ProtoBrainMaps: Prototypical Brain Maps for Alzheimer's Disease Progression Modeling - (short)
• F8: Nele Blum - Projection Domain Metal Artifact Reduction in Computed Tomography using Conditional Generative Adversarial Networks - (short)
• F9: Attila Tibor Simko - Changing the Contrast of Magnetic Resonance Imaging Signals using Deep Learning - (full)

Study Group 4-E

• Mentor: Sriprabha Ramanarayanan (1+5)
• Session Chairs G: Angelica Aviles-Rivero, Antoine Théberge, Chairs H: Raghav Mehta, Clarisa Sanchez

This study group discusses the following papers

• G1: Guanghui FU - Me-NDT: Neural-backed Decision Tree for Visual Explainability of Deep Medical Models - (short)
• G2: Cher Bass - ICAM-reg: Interpretable Classification and Regression with Feature Attribution for Mapping Neurological Phenotypes in Individual Scans - (short)
• G3: Lasse Hansen - Radiographic Assessment of CVC Malpositioning: How can AI best support clinicians? - (short)
• H4: Brennan Nichyporuk - Optimizing Operating Points for High Performance Lesion Detection and Segmentation Using Lesion Size Reweighting - (short)
• H5: Junyu Chen - Creating Anthropomorphic Phantoms via Unsupervised Convolutional Neural Networks - (short)
• H12: Siyu Shi - Unseen Disease Detection for Deep Learning Interpretation of Chest X-rays - (full)

Study Groups Thursday 23:45-0:30 (UTC+2)

Study Group 5-A

• Mentor: Maria Vakalopoulou (3+1)
• Session Chairs: Vakalopoulou

This study group discusses the following papers

• I7: Elisabeth Sarah Lane - Echocardiographic Phase Detection Using Neural Networks - (short)
• I8: Saverio Vadacchino - HAD-Net: A Hierarchical Adversarial Knowledge Distillation Network for Improved Enhanced Tumour Segmentation Without Post-Contrast Images - (full)
• I9: Mikael Brudfors - An MRF-UNet Product of Experts for Image Segmentation - (full)
• I10: Fahdi Kanavati - Partial transfusion: on the expressive influence of trainable batch norm parameters for transfer learning - (full)

Study Group 5-B

• Mentor: Tal Arbel (3+2)
• Session Chairs L: Geert Litjens, Ozan Oktay Chairs K: Hoel Kervadec, Max-Heinrich Laves

This study group discusses the following papers

• L1: Louis van Harten - Untangling the Small Intestine in 3D cine-MRI using Deep Stochastic Tracking
• L2: Niharika Shimona Dsouza - M-GCN: A Multimodal Graph Convolutional Network to Integrate Functional and Structural Connectomics Data to Predict Multidimensional Phenotypic Characterizations
• L3: JUAN LIU - Improved model-based deep learning for quantitative susceptibility mapping
• K1: Niamh Belton - Semi-Supervised Siamese Network for Identifying Bad Data in Medical Imaging Datasets - (short)
• K2: Joshua Thomas Bridge - mGEV: Extension of the GEV Activation to Multiclass Classification - (short)

Study Group 5-C

• Mentor: Hervé Lombaert (3+0)
• Session Chairs: Ismail Ben Ayed, Arrate Muñoz-Barrutia

This study group discusses the following papers

• I1: Caner Ozer - Explainable Image Quality Analysis of Chest X-Rays
• I2: Joseph Paul Cohen - Gifsplanation via Latent Shift: A Simple Autoencoder Approach to Counterfactual Generation for Chest X-rays
• I3: Abhejit Rajagopal - Understanding and Visualizing Generalization UNets

Study Groups Friday 11:45-12:30 (UTC+2)

Study Group 6-A

• Mentor: Katharina Breininger (2+3)
• Session Chairs: Katharina Breininger, Maria Vakalopoulou

This study group discusses the following papers

• I4: Yassine Barhoumi - Scopeformer: n-CNN-ViT hybrid model for Intracranial hemorrhage subtypes classification - (short)
• I5: Neerav Karani - Robust medical image segmentation by adapting neural networks for each test image - (short)
• I6: Marc Aubreville - Quantifying the Scanner-Induced Domain Gap in Mitosis Detection - (short)
• I11: Sobhan Hemati - CNN and Deep Sets for End-to-End Whole Slide Image Representation Learning - (full)
• I12: Khrystyna Faryna - Tailoring automated data augmentation to H&E-stained histopathology - (full)

Study Group 6-B

• Mentor: Nikolas Lessmann (3+2)
• Session Chairs: Nikolas Lessmann, Nick Pawlowski

This study group discusses the following papers

• J1: Hendrik J. Klug - Multimodal Generative Learning on the MIMIC-CXR Database - (short)
• J2: Simeon Emilov Spasov - TG-DGM: Clustering Brain Activity using a Temporal Graph Deep Generative Model - (short)
• J5: Jannis Hagenah - Discrete Pseudohealthy Synthesis: Aortic Root Shape Typification and Type Classification with Pathological Prior - (full)
• J6: Matteo Dunnhofer - Improving MRI-based Knee Disorder Diagnosis with Pyramidal Feature Details - (full)
• J7: Umang Gupta - Membership Inference Attacks on Deep Regression Models for Neuroimaging - (full)

Study Group 6-C

• Mentor: Nick Pawlowski (2+2)
• Session Chairs: Nikolas Lessmann, Nick Pawlowski

This study group discusses the following papers

• J3: Daniel Wulff - Comparison of Representation Learning Techniques for Tracking in time resolved 3D Ultrasound - (short)
• J4: Weiyi Xie - Deep Clustering Activation Maps for Emphysema Subtyping - (short)
• J8: Hristina Uzunova - Guided Filter Regularization for Improved Disentanglement of Shape and Appearance in Diffeomorphic Autoencoders - (full)
• J9: Camila Gonzalez - Self-supervised Out-of-distribution Detection for Cardiac CMR Segmentation - (full)

Study Group 6-D

• Mentor: Arrate Muñoz-Barrutia (3+0)
• Session Chairs: Ismail Ben Ayed, Arrate Muñoz-Barrutia

This study group discusses the following papers

• I1: Caner Ozer - Explainable Image Quality Analysis of Chest X-Rays
• I2: Joseph Paul Cohen - Gifsplanation via Latent Shift: A Simple Autoencoder Approach to Counterfactual Generation for Chest X-rays
• I3: Abhejit Rajagopal - Understanding and Visualizing Generalization UNets

Study Group 6-E

• Mentor: Ismail Ben Ayed (3+1)
• Session Chairs K: Hoel Kervadec, Max-Heinrich Laves

This study group discusses the following papers

• K3: Sonia Martinot - Weakly supervised 3D ConvLSTMs for high precision Monte-Carlo radiotherapy dose simulations - (short)
• K7: Lennart Bargsten - Attention via Scattering Transforms for Segmentation of Small Intravascular Ultrasound Data Sets - (full)
• K8: Raouf Muhamedrahimov - Learning Interclass Relations for Intravenous Contrast Phase Classification in CT - (full)
• K9: Tianshu Chu - Improving Weakly Supervised Lesion Segmentation using Multi-Task Learning - (full)