Skip to main content

Preparing for power outages at UEF

Text stickers on glass wall.

Doctoral defence of Juan Miguel Valverde, MSc, 14 Oct 2022: Brain MRI segmentation with deep learning

The doctoral dissertation in the field of Medical Image Analysis will be examined at the Faculty of Health Sciences. The public examination will be streamed online.

What is the topic of your doctoral research? Why is it important to study the topic?

My doctoral dissertation focuses on the segmentation of magnetic resonance images with deep learning. This is important because segmentations are required for conducting quantitative analyses and segmenting these images manually is tedious. I developed and applied new segmentation methods on large preclinical magnetic resonance imaging data, and I investigated strategies to increase the applicability of these methods in heterogeneous datasets.

What are the key findings or observations of your doctoral research?

The segmentations produced with the developed methods were of higher quality than with other methods, and these automatic segmentations were more accurate than the segmentations annotated by an expert, which is important for reproducibility.

How can the results of your doctoral research be utilised in practice?

The methods presented in this dissertation are open source and accessible to any researcher. Furthermore, the categorization of strategies to tackle heterogeneous datasets revealed research trends and gaps, which can aid in future investigations.

What are the key research methods and materials used in your doctoral research?

I developed three methods for segmenting brain MRI volumes of rodents with lesions: two of them focused on lesion segmentation, and the third focused on cerebral hemisphere segmentation. These methods were based on deep learning, which is a predominant approach in medical image segmentation. I also developed a mathematical function that aids in creating segmentation models for any type of medical imaging.

The doctoral dissertation of Juan Miguel Valverde, Master of Science, entitled MRI Segmentation with Deep Learning. Application in Rodent Brain Images with Lesions will be examined at the Faculty of Health Sciences. The Opponent in the public examination will be Professor Koen Van Leemput of the Technical University of Denmark, and the Custos will be Professor Jussi Tohka of the University of Eastern Finland. The public examination will be held in English.

Public examination

Downloadable photo

Doctoral dissertation

For further information, please contact:

Juan Miguel Valverde, juanmiguel.valverde (a),