ICML Workshop on Machine Learning meets Medical Imaging

The 1st ICML workshop on Machine Learning meets Medical Imaging will be held on 11th July 2015 in Lille, in conjunction with the International Conference on Machine Learning.

This workshop will present original methods and applications on Machine Learning in Medical Imaging. Developments in machine learning have opened up a wealth of novel opportunities in knowledge discovery, analysis, visualisation and reconstruction of medical image datasets. However, medical images also pose several particular challenges for standard approaches, for instance, lack of data availability (due to ethics or rarity of pathology), poor image quality (due to imaging or medical condition) or dedicated training requirements.

This workshop offers a unique opportunity to present and discuss their latest work on Machine Learning in Medical Imaging in the presence of both machine learning and medical imaging communities.  Innovative contributions will address questions such as how to better exploit smaller datasets, and understand fundamentals on image spaces or on generative models in order to improve training of machine learning methods.  The workshop will focus on theoretical aspects as well as on effective applications built on machine learning and all aspects of medical imaging.

Our Sponsors:

Best Paper Award sponsored by:

Topic of Interests

We invite innovative contributions in all aspects of machine learning in medical imaging, including:

  • Medical image analysis (segmentation, labelling, registration)
  • Modeling spaces of medical images  
  • Learning on datasets of medical data
  • Retrieval of medical images
  • Classification of medical images
  • Dimensionality reduction
  • Sparse methods
  • Medical image reconstruction (CT, PET, MRI)

Invited Speakers

  • Bertrand Thirion  Research Director @INRIA France
  • John Ashburner  Professor @Functional Imaging Laboratory, University College London, UK
  • Marleen de Bruijne  Professor @Erasmus Medical Center, Rotterdam, NL and @DIKU, CS Dept., Copenhagen, DK
  • Ben Glocker  Lecturer @Imperial College London, UK

In addition, a representative from NVIDIA will be giving a talk on the "NVIDIA GPU platform for deep learning".

Best Paper Award

Congratulations to Jonathan Young, for his paper "Improving MRI brain image classification with anatomical regional kernels"

** Programme - July 11th, 8:45-18:00 **

The programme is provisional and subject to change:

8.45    Open
9.00    Julien Demouth, NVIDIA
9.15    Invited talk 1:  Ben Glocker  Lecturer @Imperial College London, UK
Semantic Imaging: Learning to Understand Medical Images

10.00-10.30   Coffee

10.30-11.00: Motion
10.30    Retrospective motion correction of magnitude-input MR images. Alexander Loktyushin
10.45    Automatic brain localisation in foetal MRI using superpixel graphs. Amir Alansary

11.00    Invited talk 2:  Bertrand Thirion  Research Director @INRIA France
Learning representations from functional brain images.

11.45-12.30: Brain 
11.45    Learning Deep Temporal Representations for fMRI Brain Decoding. Orhan Firat
12.00    Modelling Non-Stationary and Non-Separable Spatio-Temporal Changes in Neurodegeneration via Gaussian Process Convolution. Marco Lorenzi
12.15    Improving MRI brain image classification with anatomical regional kernels. Jonathan Young

12.30-14.30    Lunch

14.30    Invited talk 3:  John Ashburner  Professor @Functional Imaging Laboratory, University College London, UK
Applying pattern recognition to anatomical scans. 

15.15-16.00 Computer Aided Diagnosis
15.15    A Graph Based Classification Method for Multiple Sclerosis Clinical Form Using Support Vector Machine. Claudio Stamile
15.30    Classification of Alzheimer’s Disease using Discriminant Manifolds of Hippocampus Shapes. Mahsa Shakeri
15.45    Transfer Learning for Prostate Cancer Mapping Based on Multicentric MR imaging databases. Rahaf Aljundi

16.00-16.30    Coffee

16.30    Invited talk 4:  Marleen de Bruijne  Professor @Erasmus Medical Center, Rotterdam, NL and @DIKU, CS Dept., Copenhagen, DK
Learning Imaging Biomarkers: Experiences From a Lung Cancer Screening Trial

17.15-17.45 Segmentation
17.15    Feature-Space Transformation Improves Supervised Segmentation Across Scanners. Annegreet van Opbroek
17.30    Discriminative Dimensionality Reduction for Patch-based Label Fusion. Gerard Sanroma

17.45    Closing + best paper award


The workshop will be held on Saturday 11th July in the Lille Grand Palais, in the room Jeanne de Flandre 2, in conjunction with the International Conference on Machine Learning 2015. Further information on registration and accommodation can be found on the ICML 2015 website.

Important Dates

  • May 3rd    Paper registration, Abstract due
  • May 4th     Final manuscript due
  • May 10th   Acceptance notification
  • May 15th   ICML Early Registration Deadline
  • July 11th    Workshop
  • July 31st    Camera-Ready Submission

Paper submission

We are pleased to announce that LNCS will be accepting camera-ready versions of up to 14 pages. That means plenty of space to improve your manuscripts, enrich your discussions with comments from the conference, reviewers, improve your figures with extra clarity.  

You can now upload your complete Camera-Ready LaTeX files on the submission website, including:
  1. Complete Revised LaTeX files (including figures)
  2. Signed Consent to Publish Form (PDF)
  3. PDF of article for checking
Accepted papers will be published as post-proceedings in LNCS format, (Latex package here), in an LNCS Springer volume. 
The best paper (judged on the day) will be awarded with an NVIDIA GPU.

Please submit your paper using the following link: https://cmt.research.microsoft.com/MEDIM2015/Default.aspx.