Automatic cartilage segmentation and measurement for diagnosis of OA using 3D Box and Gaussian filters

초록

This paper presents an automatic cartilage segmentation using 3D Box filter and multi-level Gaussian filters for diagnosis of Osteoarthritis (OA) with gradient multi-echo Magnetic Resonance Images (MRI). The 3D Box filter is applied for automatically detecting the Volume of Interest (VOI) of bone, and the multi-level Gaussian filters are applied for accurately computing the 3D normal vectors of cartilage. Then, the measurement of cartilage thickness and volume is made along the direction of the normal vectors in the VOI. Experimental results demonstrate the robustness and efficiency of the proposed approach to 3D cartilage segmentation in the gradient multi-echo MRI.