Key dates

  • 1 July
  • Abstract registration
  • 16 July
  • Paper submission
  • 6 August
  • Notification of acceptance
  • 15 August
  • Camera ready paper
  • 13 October
  • Workshop

Program

8:00-8:35 Welcome
8:35-9:15 Disentangled representation learning in medical imaging
Keynote speaker: Professor Sotirios A. Tsaftaris (University of Edinburgh, UK)

image-leftThe detection of disease, segmentation of anatomy and other classical analysis tasks, have seen incredible improvements due to deep learning. Yet these advances need lots of data: for every new task, new modality, new hospital more training data are needed. In this talk, I will show how deep neural networks can learn latent and disentangled embeddings suitable for several analysis tasks in the heart. Within a multi-task learning setting I will show that the same framework can learn embeddings drawing supervision from self-supervised tasks that use reconstruction and also temporal dynamics overall reducing considerably need for annotation. I will discuss how different architectural choices can solve key problems in multimodal data processing and critically also allow us to learn reducing need for image-level annotations by obtaining supervision also from text-based health records. I will conclude by highlighting challenges for deep learning in healthcare in general.
9:15-10:30 Regular papers
9:15-9:27 Assessing the Impact of Blood Pressure on Cardiac Function Using Interpretable Biomarkers and Variational Autoencoders (EP Anton, B Ruijsink, J Clough, I Oksuz, D Rueckert, R Razavi, A King)
9:27-9:40 A Cascade Regression Model for Anatomical Landmark Detection (Z Tan, Y Duan, Z Wu, J Feng, J Zhou)
9:40-9:52 Comparison of 2D Echocardiography and Cardiac Cine MRI in the Assessment of Regional Left Ventricular Wall Thickness (V van Hal, D Zhao, K Gilbert, TPB Gamage, C Mauger, RN Doughty, ME Legget, J Zhao, A Nalar, O Camara, AA Young, VY Wang, M Nash)
9:52-10:05 Learning interactions between cardiac shape and deformation: application to pulmonary hypertension (M Di Folco, P Clarysse, P Moceri, N Duchateau​)
10:05-10:17 Deep Learning Surrogate of Computational Fluid Dynamics for Thrombus Formation Risk in the Left Atrial Appendage (X Morales, J Mill, KA Juhl, A Olivares, G Jimenez-Perez, RR Paulsen, O Camara)
10:17-10:30 Probabilistic Motion Modeling from Medical Image Sequences: Application to Cardiac Cine-MRI (J Krebs, T Mansi, N Ayache, H Delingette)
10:30-10:45 Coffee break
10:45-12:00 CRT-EPIGGY19 Challenge
10:45-11:00 Best (and worst) practices for organizing a challenge on cardiac biophysical models during AI summer: the CRT-EPiggy19 challenge (O Camara)
11:00-11:10 Electrophysiological Model Personalisation to Porcine in-vivo Data for Paced Activation Prediction (​Cedilnik, M Sermesant)
11:10-11:20 Evaluation of meshless methods for cardiac electrophysiological simulations based on porcine experimental data (KA Mountris, E Pueyo​)
11:20-11:30 Prediction of electrical activation patterns after cardiac resynchronization therapy in porcine hearts with meshless models (C Albors, E Lluch, JM, R Doste, O Camara, M De Craene, H Morales​)
11:30-11:40 Prediction of CRT activation sequence by personalization of biventricular model from electroanatomical maps (JF Gomez, B Trenor, R Sebastian)
11:40-11:50 Optimization of CRT therapy device based on personalized computer model (S Khamzin, A Dokuchaev, O Solovyova)
11:50-12:00 Challenge wrap-up
12:00-12:36 Regular poster teaser (2 min for each poster
12:36-14:00 Lunch, poster presentation, scoring
14:00-15:15 LV Full Quantification Challenge
14:00-14:20 Challenge data and description from the organizers
14:20:14:30 Left Ventricle Quantification Using Direct Regression with Segmentation Regularization and Ensembles of Pretrained 2D and 3D CNNs (N Gessert, A Schlaefer)
14:30-14:40 Left Ventricle Quantification with Cardiac MRI : Deep Learning Meets Statistical Models of Deformation (JC Acero, H Xu, E Zacur, J Schneider, P Lamata, A Bueno-Orovio, V Grau​)
14:40-14:50 Left Ventricular Parameter Regression From Deep Feature Maps of a Jointly Trained Segmentation CNN (S Tilborghs, F Maes​​)
14:50-15:00 A Two-Stages Temporal-Like Fully Convolutional Network Framework for Left Ventricle Segmentation (Z Zhao, N Boutry, E Puybareau, T Géraud)
15:00-15:15 Challenge wrap-up
15:15-15:30 Coffee break
15:30-17:00 Multi-Sequence Cardiac MR Segmentation Challenge
15:30-15:40 Challenge introduction
15:40-15:53 Unsupervised Multi-modal Style Transfer for Cardiac MR Segmentation (C Chen, C Ouyang, G Tarroni, J Schlemper, H Qiu, W Bai, D Rueckert​)
15:53-16:06 Combining Multi-Sequence and Synthetic Images for Improved Segmentation of Late Gadolinium Enhanced Cardiac MRI (VM Campello, C Martin-Isla, CI Morcillo, MAG Ballester, K Lekadir​)
16:06-16:19 Multi-sequence Cardiac MR Segmentation with Adversarial Domain Adaptation Network (J Wang, H Huang, C Chen, W Ma, Y Huang, X Ding​)
16:19-16:32 SK-Unet: an Improved U-net Model with Selective Kernel for the Segmentation of Multi-sequence Cardiac MR​ (X Wang, S Yang, M Tang, Y Wei, X Han, L He, J Zhang​)
16:32-16:45 Automated Multi-sequence Cardiac MRI Segmentation Using Supervised Domain Adaptation​ (S Vesal, N Ravikumar, A Maier​)
16:45-16:58 Cardiac Segmentation of LGE MRI with Noisy Labels (H Roth, W Zhu, D Yang, Z Xu, D Xu​)
17:00-17:15 Closing and awards