Dr. Veronika Cheplygina
Dr. Veronika Cheplygina, Assistant Professor at the research group Medical Image Analysis of the department of Biomedical Engineering at Eindhoven University of Technology, will present her lecture on "Multiple instance learning in biomedical applications."
Meet the speaker
Machine learning has vast potential in medical image analysis, improving possibilities for early diagnosis and prognosis of disease. However, machine learning algorithms need 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. To address these problems, several approaches have been proposed. In this talk I will discuss one particular type of approach, multiple instance learning, which is aimed adapting the algorithms to use other types of annotated data, that is available at less cost. After a brief introduction to multiple instance learning, I will review some example applications in medical imaging, a few examples from bioinformatics, and some similarities and differences between the two. I will close with discussing some open problems with these approaches.
Børge Solli Andreassen, PhD student at the Research Group for Digital Signal Processing and Image Analysis, will present his work on "Mitral Annulus Segmentation using Deep Learning in 3D Transesophageal Echocardiography."