Automatic segmentation of articular cartilage from combined assessment of sodium and proton MR knee images

Articular cartilage (AC) segmentation from magnetic resonance images of knee joint play a major role in quantifying morphological and physiological changes during the onset of osteoarthritis (OA). Prior research in cartilage segmentation has shown difficulties to develop fully automatic method due to heterogeneous AC signal, low tissue contrast, uneven shape and complex behaviour. This work aims to develop an automated AC segmentation technique utilizing combined assessment of sodium and proton MR images acquired at 1.5T.