The header of the Tilburg center for Cognition and Communication

Tilburg center for Cognition and Communication (TiCC)

We study how people communicate with each other and how computer systems can be taught to communicate with us.


TiCC Colloquium: Veronika Cheplygina

What: Learning with less labels in medical image analysis
Where: AZ 211
When: Wednesday, 20 December 2017, 12:45 - 13:45 hours


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, and discuss examples from my own work on classifying chronic obstructive pulmonary disease (COPD) in chest CT images.

About Veronika Cheplygina

Veronika Cheplygina is an assistant professor at the Medical Image Analysis group, Eindhoven University of Technology. She received her PhD 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

When: 20 December 2017 12:45

End date: 20 December 2017 13:45