Bone shape characterization using the fourier transform and edge detection in digital X-ray images

From the series of digital X-ray images we extracted bone edges using information entropy based edge detection algorithms. The extracted edges are series of two dimensional coordinates from image space. The space series were tested for shapes of medical interest inside the knee joint. In particular, the existence of straight edges inside the knee joint was tested for by using digital filtering and analytic computation of first and second derivative in the Fourier domain. This kind of analysis was used because it provides rotational invariance for the sought bone shape, and allows the statistical comparison of shapes from different images. The real life examples are taken from medical practice using X-Ray imaging of series of knee joints in order to illustrate the analysis procedures and their medical relevance for real data.