Segmentační model pro automatizovanou analýzu dynamiky periostálního svalku z RTG obrazů

Abstract

The aim of this diploma thesis is to design a segmental model of the periosteal callus with subsequent analysis of geometric parameters and density dynamics during fracture treatment. X-ray images of bone fractures were used in this work. The analysis was performed using variable bones, namely the femur, clavicle and lower leg bones (tibia and fibula). The whole segmentation process was applied in the MATLAB environment. The algorithm consists of a procedure that includes region of interest determination, contrast and brightness adjustment, image filtering, Fuzzy C means, active contours, and morphological operations. The most accurate segmentation results occured on a periosteal muscle formed on the femur. Due to the diversity of periosteal muscle formation and size during bone healing, it is necessary to specify the segmentation settings for more accurate results. The development of the periosteal muscle during the treatment was performed using indicators of geometric parameters and density dynamics, which were represented by the brightness spectrum of the image. Geometric parameters show the development of the periosteal muscle over time. The analysis of the results of the density dynamics shows an increase in the density of the muscle during the observation of the fracture treatment. The analysis shows an identification of periosteal muscle density relative to healthy bone at the time of treatment.