Automated segmentation of knee cartilage and meniscus using convolutional neural networks

PhD study program: Measurement Technology
Akademic year: 2019-020
Advisor: Mgr. Vladimír Juráš, PhD. (umervjur@savba.sk)
External educational institution: Institute of Measurement Science SAS
Accepting university: Faculty of Electrical Engineering and Information Technology, Slovak University of Technology in Bratislava, Institute of Electrical Engineering
Max. number of students: 1
Study language: Slovak/English

Annotation:
The determination of prognosis and treatment response in osteoarthritis (OA) is an important task for musculoskeletal magnetic resonance (MRI). Early stages of OA are accompanied by glycosaminoglycan loss and depletion of collagen matrix. Quantitative MRI may help in detecting macromolecular content in cartilage and meniscus in the future. Since manual segmentation of knee cartilage is a tedious task, automated approaches become more and more popular in the last years. Convolutional neural networks (CNN) may provide fast and robust solution. PhD student will be responsible for designing the neural network architecture, training on the dataset either gathered from OAI database or acquired at modern MRI scanners at Institute of Measurement Science or Medical University of Vienna. The requirements for successful thesis is the passion for programming (Matlab, Python) and for biology/medicine.

back