Veronika Cheplygina wins Lorentz-eScience competition 2018

Veronika Cheplygina won the second Lorentz-eScience competition and together with her partners received funding to support a 5-day workshop for up to 25 people in “Crowdsourcing for Medical Image Analysis”. Crowdsourcing is the process of outsourcing tasks to internet users, often with the goal of collecting data to improve machine learning techniques. An example is tagging photos on Facebook, which helps to improve face recognition software. The workshop will take place in July 2018.

The workshop will bring together researchers working on machine learning techniques for medical images, where it is very time-consuming to collect labeled data. Instead of tagging people in photos, users could be presented with a slice from a CT or MR scan, or a microscopy image, and tag parts of the image corresponding to example patterns. Although this seems difficult without medical training, several studies have shown promising results, especially if the (noisy) tags from the crowd are combined. The workshop will serve as a venue to explore this exciting topic further.

The Lorentz-eScience competition is an initiative of the Netherlands eScience Center and the Lorentz Center; the competition aims to host a leading-edge workshop on digitally enhanced research, and bring together researchers from the academic scientific community and the public/private sector. For this proposal, Veronika Cheplygina collaborated with Lora Aroyo (VU University Amsterdam), Alessandro Bozzon (Delft University of Technology), Danna Gurari (University of Texas at Austin, USA) and Zoltán Szlavik (IBM Center for Advanced Studies Benelux). The participants in the workshop are specialists from over the world in e.g. medical imaging, crowdsourcing and medical visualization.

Veronika Cheplygina is an assistant professor at the Medical Image Analysis group in the department of Biomedical Engineering, Eindhoven University of Technology. In 2015 she received her Ph.D. from the Delft University of Technology for her thesis ``Dissimilarity-Based Multiple Instance Learning“. From 2015 to 2016 she was a postdoc at the Biomedical Imaging Group Rotterdam, Erasmus Medical Center, 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.