Development and validation of algorithms for segmentation of lower extremity structures from magnetic resonance images
PhD Fellow Yunsub Jung
To fully understand knee osteoarthritis (KOA) pathogenesis and progression, it is important to understand the biomechanical environment of anatomically complex joint structures. However, since it is a difficult task to do this experimentally, testing the underlying joint mechanism through a musculoskeletal model is necessary. Musculoskeletal and finite element joint models can help estimate the progression of osteoarthritis for each patient, measure the effectiveness of a specific treatment and estimate the appropriate time for a patient's surgical intervention.
In particular, biomechanical analysis using a subject-specific musculoskeletal computational model can find a personalised treatment and management method. It uses a patient’s actual anatomical information from medical imaging. Subject-specific musculoskeletal modelling requires structural (anatomical) information about the human body. For this, manual segmentation is usually performed, which is a task that requires a lot of time and human resources.

Therefore, this project aims to develop novel algorithms to automatically segment the musculoskeletal structure (bone, cartilage, ligament, etc.) of the lower extremities from magnetic resonance imaging (MRI) images. This project includes the development of segmentation algorithms and studies on correcting irregular signals of MRI images and selecting MRI sequences for the best segmentation performance. The ultimate goal of this project – the result of automatic segmentation of lower extremity musculoskeletal structures – will be essential in progressing the field of patient-specific modelling of knee OA patients.
Yunsub holds a Master of Medicine degree from the Department of biomedical engineering at Seoul National University, South Korea. He has worked in Samsung Electronics and GE Healthcare for 15 years. You can see more information about Yunsub on Aalborg University’s homepage.