An improved algorithm for skeletonisation of nailfold capillaries

Nailfold capillaroscopy (NC) is a medical imaging modality used to assess the characteristics and morphology of micro-blood vessels in the nailfold. NC is of particular importance in diagnosing diseases that lead to morphological changes of capillaries such as scleroderma, Raynaud's phenomenon and other connective tissue diseases. In order to provide a computer-aided diagnosis approach to analysing NC images, in this paper, we present a skeletonisation algorithm of captured nailfold capillaries. Our approach first enhances the image using an a-trimmed filter, and then performs a binarisation step based on difference of Gaussian filtering and thresholding, followed by iterative thinning to extract the capillary skeletons. We show that the proposed algorithm works well and gives significantly superior performance compared to previous approaches.