Parameterisation of human femur bone using computerised tomography scan images

Abstract

In this article, we have re-constructed the human femur from computed tomography images through an advanced image processing package MIMICS®. The computerised tomography images of the human lower limb in DICOM format were imported to this package and was optimised after re-defining the images in bone scale (−1024 to +1650 HU). Masks were created, thresholding the slices in desired scale (+226 to +1,650 HU for cortical bone) to get the desired region of interest and the 3D computerised model. The computerised tomography slices were also used to create the bounding area of the femur cross-sections at various locations as recorded by the computerized tomography scanner. The centre of mass of each of the sections were determined using the pre-processor of ANSYS®. The variation of abscissa and ordinates were plotted against the length from proximal to distal end, which are closer to circles except the two extremities. Observing the circular nature of the centre of mass plots, we propose to fit the non-linearity of the femoral stem in the form of an arc of a circle. Few parameters are introduced along with existing parameters to define the femoral geometry. This will help in the process of hip prosthesis design.