Image analysis of nailfold capillary patterns from video sequences

Abstract

Video capillaroscopy is a widely used technique to assess the condition of the blood capillaries in the nailfold of patients suffering from reduced peripheral circulation (Raynaud’s phenomenon). This approach is complicated by the fact that gaps in the flow of blood can render capillaries incomplete in any single video frame. This paper describes a computer based system in which sequences of video frames are registered and combined to provide a composite image for analysis. We show that the images can be registered to an accuracy of approximately 2 μm, and that temporally variable features can be enhanced significantly by subtracting one standard deviation from the mean value for each pixel position in the video sequence.