Speakers

External speakers and contributors

Christophe Avare, Research and Innovation Director Olea Medical

 

Christophe Avare, PhD, has joined the Research & Innovation Department to help develop new approach and products leveraging data analytics and deep learning. Over the last 20 years he has been involved in many large-scale distributed architecture designs for space, aeronautics and transportation solutions as a Senior Software Architect for Dassault Systèmes, then the Thales Group, but always with a focus on challenges for data analysis. From 2012 to 2016, he worked in the Reasoning and Analysis in Complex Systems Laboratory of the French Thales Research and Technology facility and was one of the Software Expert nominated to the Patent committee. He was also the Big Data, Web and Cloud Technologies Segment Manager for the Thales Group and has been involved in the strategy and research program definitions in key technology domains like Cybersecurity, Predictive Maintenance and Earth Observation. Given the renewed interest in what is now sometimes abusively called “AI”, Christophe will help build bridges between these technologies and application domains to tackle future challenges in the next generation of medical image post-processing products. LinkedIn: https://fr.linkedin.com/in/avare

Veronika Cheplygina, Eindhoven University of Technology, The Netherlands

 Veronika Cheplygina is an assistant professor at the Medical Image Analysis group, Eindhoven University of Technology since February 2017. She received her Ph.D. from the Delft University of Technology for her thesis ``Dissimilarity-Based Multiple Instance Learning“ in 2015. As part of her PhD, she was a visiting researcher at the Max Planck Institute for Intelligent Systems in Tuebingen, Germany. From 2015 to 2016 she was a postdoc at the Biomedical Imaging Group Rotterdam, Erasmus MC, where she applied machine learning algorithms to medical image analysis problems. Her research interests are centered around learning scenarios where few labels are available, such as multiple instance learning, transfer learning, and crowdsourcing. Next to research, Veronika blogs about academic life at  http://www.veronikach.com     

Abstract:

Learning with less labels in medical image analysis

Machine learning (ML) has vast potential in medical image analysis, improving possibilities for early diagnosis and prognosis of disease. However, ML needs large amounts of representative, annotated examples for good performance. The annotation process, often consisting of outlining structures in (possibly 3D) medical images, is time-consuming and expensive. Furthermore, annotated data may not always be representative of new data being acquired, for example due to changes in scanners and scanning protocols. In this talk I will give an overview of approaches such as multiple instance learning and transfer learning, used to address these challenges, discuss their connections and several (underexplored) directions for future research.

Paul De Brem, journaliste scientifique, Paris

 

Paul de Brem is a professional anchorman specialised in space, scientific and technical events. During the last 15 years, he has hosted more than 500 symposiums and debates for various clients such as CNRS, Procter&Gamble, Région Ile-de-France, Inserm, EDF, Sanofi, Institut Pasteur, ministère de la Recherche, CNES, etc.

A year ago, he has hosted a two-day ministerial conference in english dedicated to higher education with 48 ministers from 4 continents.

He also leads communication courses for professionals: Media-training, Powerful PowerPoint, Writing for the Internet, etc. for clients such as CNES, Banque de France, Orange, the ENA (Ecole nationale d’administration), etc.

He has been leading courses in scientific journalism at Sorbonne Université for

8 years. Previously, as a science editor for television and printed media, he actively collaborated with LCI, France 2, France 24, le Journal du dimanche, L’Express, etc.

Christian Desrosiers, École de Technologie Supérieure, Canada

Prof. Desrosiers obtained a Ph.D. in Applied Mathematics from Polytechnique Montreal in 2008, and was a postdoctoral researcher at the University of Minnesota with prof George Karypis. In 2009, he joined École de technologie supérieure (ÉTS) as professor in the Departement of Software and IT Engineering. He is codirector of the Laboratoire d’imagerie, de vision et d’intelligence artificielle (LIVIA) and a member of the REPARTI research network. He has over 100 publications in the fields of machine learning, image processing, computer vision and medical imaging, and has served on the scientific committee of several important conferences in these fields.

Abstract (common lecture with Pierre-Marc Jodoin):

Basics in deep learning, part 1 and 2

In these two lectures, we will address the fundamentals of neural networks.  We will revise how the first neural network (aka the "Perceptron") was used as a machine for linear classification and regression and how it relates to logistic regression.  We will then show how one can train it with well-known Stochastic Gradient Descent that minimizes a loss function such as L2, Cross-entropy, and KL-Divergence.  We will also show how to make the Percepton become "deep" and multi-class by adding multiple layers and activation functions to it.  Other key topics will be addressed such as "forward" and "backward" propagation, learning, and mini-batches.  We will finally show what prevents the multi-layer perceptron from scaling and how it can be simplified to a Convolution neural nets (CNN).

Advanced concepts in deep learning

This lecture is the follow up of the "Basics in deep learning" lectures and will focus on state-of-the-art (SotA) CNNs.  We will present SotA structures for classification (namely ResNet, VGGNet, InceptionNet, and DenseNet ) and for segmentation (namely Encoder-decoder, UNet, VNet, SegNet, pyramidNet, and Fully-convolutional nets).  And if the time permits, we will introduce localization networks as well as recurrent neural networks (such as GRU and LSTM).

Benoit Gallix, IHU, University of Strasbourg

Prof. Benoit Gallix, MD, PhD, is former chair of the diagnostic radiology of the Department of Diagnostic Radiology, McGill University Health Center, Montreal, Canada and designated CEO of the IHU, University of Strasbourg, is a pioneer in the development of self-learning algorithms for the identification of imaging-based biomarkers of liver disease and HCC and world-class expert in hepatology. His focus is on abdominal imaging, both diagnostic and interventional, with a special interest in the therapy of cancer. When he arrived at McGill in 2013, as Chair and Director of the Imaging Department, Dr. Gallix created – in collaboration with the teams of computer Sciences from McGill (Center for Intelligent Machine) – a research program focus on Artificial Intelligence (AI). His research activities are at the interface between Oncology, Medical Imaging, and computer vision with the objective of developing new methods of tumor quantification by imaging, in order to select patients who are likely to respond to a specific treatment and to evaluate their response very early.

Pierre-Marc Jodoin, University of Sherbrooke, Canada

Pierre-Marc Jodoin is from  the University of Sherbrooke, Canada where he works as a full professor since 2007.  He specializes in the development of novel techniques for machine learning and deep learning applied to computer vision and medical imaging.   He mostly works in video analytics and brain and cardiac image analytics.  He is the co-director of the Sherbrooke AI plateform and co-founder of the medical imaging company called "Imeka.ca" which specializes in MRI brain image analytics. web site: http://info.usherbrooke.ca/pmjodoin/

Abstract (common lecture with Christian Desrosiers):

Basics in deep learning, part 1 and 2

In these two lectures, we will address the fundamentals of neural networks.  We will revise how the first neural network (aka the "Perceptron") was used as a machine for linear classification and regression and how it relates to logistic regression.  We will then show how one can train it with well-known Stochastic Gradient Descent that minimizes a loss function such as L2, Cross-entropy, and KL-Divergence.  We will also show how to make the Percepton become "deep" and multi-class by adding multiple layers and activation functions to it.  Other key topics will be addressed such as "forward" and "backward" propagation, learning, and mini-batches.  We will finally show what prevents the multi-layer perceptron from scaling and how it can be simplified to a Convolution neural nets (CNN).

Advanced concepts in deep learning

This lecture is the follow up of the "Basics in deep learning" lectures and will focus on state-of-the-art (SotA) CNNs.  We will present SotA structures for classification (namely ResNet, VGGNet, InceptionNet, and DenseNet ) and for segmentation (namely Encoder-decoder, UNet, VNet, SegNet, pyramidNet, and Fully-convolutional nets).  And if the time permits, we will introduce localization networks as well as recurrent neural networks (such as GRU and LSTM).

Anirban Mukhopadhyay, Technische Universität Darmstadt, Germany

Dr. Anirban Mukhopadhyay obtained his PhD in Computer Science with a minor in Statistics from the University of Georgia, USA in 2014. He is currently leading the junior research group Medical and Environmental Computing (MECLab) at Technische Universität Darmstadt, Germany. His current research focus is on safe translation of AI toward healthcare. He has recently spearheaded the first collaborative review on Generative Adversarial Networks for Medical Image Analysis. His research efforts have been awarded with multiple prestigious awards, including the best thesis award in mathematical sciences 2014 from the University of Georgia and a Miccai society travel award. 

https://sites.google.com/site/geometricanirban/

Abstract

 

Generative adversarial networks for medical imaging

 

Generative adversarial networks (GANs) are a powerful subclass of deep generative models that are currently receiving widespread attention from the medical imaging community. The key idea behind GANs is that two neural networks are jointly optimized in a competitive fashion: one network tries to synthesize samples that resemble real data points while a second network assesses how well the result corresponds to a reference database of samples. Adversarial methods have been successfully exploited in typical medical image analysis applications such as denoising, synthesis, reconstruction, segmentation, and detection. Moreover, adversarial training has led to new applications in paradigms such as semi-supervised learning and abnormality detection. In this lecture, I will provide basic as well as advanced material on GANs and adversarial methods in medical image analysis. We will focus on key state-of-the-art papers in the machine learning and computer vision literature and their relation to works in medical image analysis. To make these concepts tangible, I will also provide examples of applications in medical imaging.

Ozan Oktay, Imperial College London, United-Kingdom

Dr. Ozan Oktay is a Research Scientist at HeartFlow Inc. (CA, USA) and has been working as a Research Fellow at Imperial College London (ICL) since early 2018. His research focuses on development of algorithms and machine learning methodologies for medical image analysis. In his early academic career, he held a Research Associate role in Computing Department at ICL, where he worked with Prof. Daniel Rueckert. During his PhD at ICL, he developed novel methodologies for medical image reconstruction and semantic image segmentation. He has made influential contributions to the medical imaging community with publications featured in book chapters (2), top-tier journals (13), and conference proceedings (31). These research contributions have been complimented by best-paper awards in top international conferences (MICCAI'13, FIMH'15). Besides his employment at ICL, he led the core development team at ThinkSono Ltd (London, UK) as a machine learning advisor. Previously, he worked in Siemens Corporate Research (NJ, USA) and ABB (Baden, CH) for two years as a researcher.
 

Website: https://www.doc.ic.ac.uk/~oo2113/

Abstract

Deep learning in cardiovascular applications

In this talk, I will present various research projects that we have been working on at Imperial College London and HeartFlow CA, USA. The first part of presentation will mainly focus on applications of machine learning techniques in cardiac image analysis tasks, including: (I) automated semantic segmentation of coronary vessel and large structures for different imaging modalities, (II) cardiac MR image reconstruction from k-space data and image resolution enhancement, and (III) learning anatomical priors in automated image analysis tasks to produce clinically meaningful results. The discussion will be mainly centred around convolutional neural networks and it will be followed by our recent work on attention gate modelling. The last part of the talk will be dedicated to a discussion on current limitations of CNN models for their large scale deployment in clinical applications. In particular, there will be special emphasis on topics such as model generalisation and model uncertainty estimation, i.e. how can we make sure that these models can be safely applied to images acquired from different sites (varying imaging quality) and patients with different pathological background? 

Dimitris Perdios, École Polytechnique Fédérale de Lausanne, Switzerland

Dimitris Perdios received the B.Sc. and M.Sc. degrees in Mechanical Engineering from the École Polytechnique Fédérale de Lausanne (EPFL), Lausanne, Switzerland, in 2011 and 2014, respectively, where he is currently pursuing the Ph.D. degree with the Signal Processing Laboratory (LTS5), under the supervision of Prof. Jean-Philippe Thiran. He was a Research Engineer with LTS5, EPFL, from 2014 to 2016. His current research focuses on developing algorithms in the context of medical ultrasound imaging, with an emphasis on inverse problem formulations, measurement process modeling, numerical simulations and deep learning.

Nicolas Thome, CNAM, Paris, France

Nicolas Thome is a full professor at Conservatoire National des Arts et Métiers (Cnam Paris).

His research interests include machine learning and deep learning for understanding low-level signals, e.g. vision, acoustics, time series, etc. He also explores solutions for multi-modal data processing, e.g. vision and language. His current application domains are essentially targeted towards healthcare or autonomous driving.  

He is involved in several French (ANR), European and international (Canada, Singapore, Brazil) collaborative research projects on deep learning.

Abstract

Weakly supervised and disentangling

In this talk, I introduce methods for training deep neural networks when only few or coarse labels are available.

I detail multiple instance learning models, the different prediction functions and the related learning schemes used in the literature. I also provide how to use these models with deep neural networks, and their application for visual recognition in medical image analysis.

I also introduce machine learning models for leveraging unlabeled data, i.e. unsupervised and semi-supervised learning. I present standard methods based on auto-encoders, recent extensions to better disentangle supervised from unsupervised signals, and self-supervised approaches. Applications for regularizing the training of deep neural networks in medical applications will be given.

 

Nicolas Villain, Head of Research in AI for Medical Imaging at Philips France

 

Since April 2018, Nicolas Villain leads the "Hub AI Paris” created by Philips in France, a research team of approximately 40 scientists focused on AI in medical imaging. The hub also hosts a development team for Philips collaborative AI platform and an activity of venture capital that invests in AI start-ups. Before this, he was the Head of Philips Research in France since 2012.

Graduated in biomedical engineering from CentraleSupelec (FR) and Ecole Polytechnique Montreal (CA), he started as a researcher in computer vision, working in various applications such as robotics and medical imaging. He published articles in journals such as IEEE TBME, IEEE TMI and many conferences. He also holds more than 10 patents in medical imaging

Local speakers and contributors

Olivier Bernard, CREATIS laboratory, Lyon, France

Dr. Olivier Bernard has an MSc in Electrical Engineering and received a PhD in Medical Image Processing from the University of Lyon (INSA) - France - in 2006. In 2007, he was a postdoctoral research fellow at the Federal Polytechnic Institute of Lausanne (Switzerland) in the laboratory headed by Prof. Michael Unser.

In 2007, he became Associate Professor at the French University of Lyon and a member of the CREATIS laboratory (CNRS 5220, INSERM U1044, INSA-Lyon, University of Lyon). In 2008, he obtained the special mention (2nd prize) for best Ph.D. in France awarded by the IEEE Engineering in Medicine and Biology Society. In September 2013, he was an invited professor at Federal Polytechnic Institute of Lausanne (Switzerland) in the laboratory headed by Prof. Jean Philippe Thiran. He was an Associate Editor for the IEEE Transactions on Image Processing Journal (2013-2016) and was a member of the technical committee of the IEEE International Conference on Image Processing and the IEEE International Symposium on Biomedical Imaging (2014).

His current research interests include medical image analysis with a particular attention to cardiac imaging. He has a strong interest in machine learning, image segmentation, motion analysis, statistical modeling, sampling theories and image reconstruction

Loïc Boussel CREATIS laboratory and Croix-Rousse Lyon Hospital (HCL), Lyon, France

 

Loic Boussel is a full professor of radiology, chairman of Department of Radiology at the Croix-Rousse Lyon Hospital (HCL). As a member of Creatis, he is involved in research in spectral CT and AI for cardio-vascular diseases.

Pierre Croisille Université Jean Monnet & CHU Saint-Etienne, CREATIS laboratory, Lyon, France

Pierre Croisille is Professor of Radiology at Université de Lyon / Université Jean Monnet (Saint-Etienne, France), and is Deputy Director of CREATIS Research Lab (CNRS 5220, INSERM U1216).  He is the Head of the Imaging Department, and Chairman of the Radiology and Nuclear Medicine in University Hospital CHU Saint-Etienne. He earned his MD and PhD degrees at the University of Lyon (France). He trained in Johns Hopkins University (Baltimore, USA) and Cantonal University Hospital (Geneva, Switzerland).  

His research is focusing on the development of innovative cardiac imaging approaches, including noninvasive new quantitative imaging methods and biomarkers to characterize myocardial and skeletal muscle damages. He is actively promoting the transfer of fundamental knowledge to the clinical needs as it is emphasized by his involvement in the development of several software solutions (inTag, CMRSegTools, CMRDiffTools) that are worldwide distributed as plugins within the open-source Horos platform in a clinical environment.

He has also experience in managing multicenter collaborative projects. He has been/is in charge of the management of the MR core-lab of several clinical trials. He is also actively involved as a board member and in charge of the supervision of a cardiac imaging data-bank (MRI, US) of the CARIM cohort that is collecting heterogeneous imaging data (raw-data, dicom files) connected to the bio-bank and clinical e-CRF, using an innovative distributed network spread across clinical sites.

He is one of the initiator of the Human Heart Project  a single point of reference to medical annotated imaging datasets that enables research teams to easily and rapidly share data, test computational methods and enhance collaboration around heart imaging and analysis (http://humanheart-project.creatis.insa-lyon.fr). Pierre Croisille is the author and/or co-author of more than 189 peer-reviewed papers mainly focusing on experimental, methodological or clinical applications of CMR (ResearchID H-4928-2014)

Nicolas Duchateau, CREATIS laboratory, Lyon, France

Nicolas Duchateau is Associate Professor (Maître de Conférences) at the Université Lyon 1 and the CREATIS lab in Lyon, France. His research focuses on the statistical analysis of medical imaging data to better understand disease apparition and evolution, and to a certain extent computer-aided diagnosis. On the technical side, it mainly covers post-processing through statistical atlases and machine learning techniques. It also includes dedicated pre-processing and validation, among which the generation of synthetic databases. On the clinical/applicative side, it covers the study of cardiac function from heart failure populations, through routine imaging data and advanced 2D/3D shape, motion and deformation descriptors.

Thomas Grenier, CREATIS laboratory, Lyon, France

 

Dr. Thomas Grenier is Associate Professor at INSA Lyon Electrical Engineering department and at the CREATIS lab in Lyon, France.

 

My research focuses on longitudinal analysis of medical data to study evolution as Multiple Sclerosis lesions, functional activity (muscle and hydrocephaly). Most of these studies involve a segmentation task and dedicated pre and post processing steps. Clustering (spatio-temporal mean-shift), semi-supervised (multi-atlas with machine learning) or fully supervised (DNN) schemes are used to solve such problems considering their specific constraints.

Rémi Emonet, LabHC, Saint-Etienne, France

Rémi Emonet is Associate Professor (Maître de Conférence) at University Jean-Monnet and is leading of the Machine Learning project at Laboratoire Hubert Curien, in Saint Étienne.
He got a Ph.D. from the Grenoble university working at Inria, and spent some years, at Idiap research institute, Switzerland, working on probablistic models for unsupervised activity modeling in videos.
His current research and contributions focus on transfer learning, deep representation learning and anomaly detection.
He likes to manipulate Bayesian approaches and to try to derive meaningful guarantees for Machine Learning algorithms.

Carole Lartizien, CREATIS laboratory, Lyon, France

Carole Lartizien received the bachelor’s degree in Nuclear Engineering from the National Polytechnic Institute, Grenoble, France, in 1996. She received the master’s degree in Biomedical Engineering and the Ph.D. degree in Image Processing from the University Paris XI, France, in 1997 and 2001, respectively. She is a Research Director of CNRS and is conducting research at the CREATIS laboratory in Lyon whose main areas of excellence concern the identification of major health issues that can be addressed by imaging and of theoretical barriers in biomedical imaging related to signal and image processing, modelling and numerical simulation. Her research interests include machine learning methods (kernel methods, deep learning) for classification problems and the prototyping of computer aided diagnostic system (CAD) for cancer and neuro- imaging.

https://www.creatis.insa-lyon.fr/~lartizien/

Sarah Leclerc, CREATIS laboratory, Lyon, France

Currently a third year PhD student at CREATIS, Sarah Leclerc is working on the automation of multi-structure cardiac segmentation in echocardiography. She investigates supervised learning approaches such as Structured Random Forests and Convolutional Neural Networks in order to provide a clinically suitable solution

Fabien Millioz CREATIS laboratory, Lyon, France

 

Fabien Millioz graduated from the École Normale Supérieure de Cachan, France and received the M.Sc. degree in 2005 and Ph.D. degree in 2009 both in signal processing from the Institut National Polytechnique of Grenoble, France. Since 2011, he is lecturer at University Claude Bernard Lyon 1, and member of the Creatis lab since 2015.

His research interests are statistical signal processing, fast acquisition, compressed sensing and neural networks.

Bruno Montcel CREATIS laboratory, Lyon, France

Bruno Montcel is Associate Professor (Maître de Conférences - HDR) at the Université Lyon 1 and the CREATIS lab in Lyon, France. His research focuses on optical imaging methods and experimental set up for the exploration of brain physiology and pathologies. It mainly focuses on intraoperative and point of care hyperspectral optical imaging methods for medical diagnosis and gesture assistance.

Michaël Sdika CREATIS laboratory, Lyon, France

Michaël Sdika is from the CREATIS lab in Lyon, France. His current research field focuses on the development of new analysis method based on deep learning for medical data. His main contributions are centered around image registration, atlas based segmentation, structure localization and machine learning for MR image of the nervous central system.

Abstract

 

Deep learning in neuroimaging

In this presentation, we will review recent works on the use of deep learning for several applications to neuroimaging. More specifically, we will present the recent advances in image registration, image segmentation, artifact correction and image enhancement based on deep neural networks.

 

 

 

Personnes connectées : 1