Three dimensional modelling of MRI knee images using improved edge detection and finite element modelling

The knee is one of the complex joints in the human body. Manual Material Handling (MMH), especially lifting, poses a risk to humans and this is considered as the preliminary reason for back pain and other impairments in a knee. Hence, it is very necessary to find the positions of knee cartilage of a person. This is usually done by modelling a three dimensional (3D) model of a patient from a Magnetic Resonant Images (MRIs). This modelling is used for determining the patterns of stress and strain distribution from various body postures. However, the modelling accuracy is not favorable in executing the required 3D patterns for finding the actual patterns of knee position. In this paper, an improved 3D model is developed for evaluating the positions of knee cartilages. The source images are captured from MRI by scanning the knee portions of a human to render a 3D knee pattern. This method works on two different stages, initially, the method uses edge-based detection and the second stage gets the edges as its input to acquire to 3D models. The second stage uses a finite element model to analyze the stresses associated with the Knee-joint and careful design is made using the proposed model to analyze the knee cartilage. The proposed model attains a good temporal agreement between the rendered 3D results and the predicted muscle force based on the magnitude of the obtained models. The proposed 3D modelling is evaluated in terms of contact mechanics factors like contact pressure, contact area and contact stress are simulated synchronously. It is found that the maximum contact pressure exists on the medial tibial insert. Finally, it is concluded from the results that the automated segmentation method to render the 3D patterns from MR images is accurate and quick in determining the geometries of knee cartilages.