Segmentation and Analysis of Knee Femoral Cartilage for Osteoarthritis using MR Images

Abstract

Abstract Knee Osteoarthritis (OA) is a chronic disease of femoral knee cartilage impairment and can be analyzed with Magnetic Resonance Imaging (MRI). Early detection of the disease helps patient from severe damages. Hence segmentation of knee cartilage is based on pixels of the image. We propose a Femoral Knee Cartilage degeneration method by 3 components: Preprocessing is done to standardize intensity and spatial characteristics and then Segmentation is carried out by registration, transformations and wrapping of reference image and results are verified using Dice Similarity Coefficients; Cartilage thickness is visualized by clouds using morphology and relaxometry. This paper aims at analyzing the femoral segmented cartilage thickness and volume via Pearson coefficients. Ground truth segmentation resulted in an average DICS Coefficient of 0.86. Pearson’s coefficients were 0.96 for cartilage thickness and 0.98 for cartilage volume.