Magers S, Faber J E
Department of Physiology, University of North Carolina, Chapel Hill 27599-7545.
J Vasc Res. 1992 May-Jun;29(3):241-7. doi: 10.1159/000158938.
A real-time automatic digital image-processing algorithm was developed for measuring microvascular dimensions. Using video microscopic images of the vasculature, the video images are digitized and analyzed using contrast enhancement, frame averaging, and pattern recognition, rather than edge detection. Performance of the algorithm was evaluated for second- and third-order vessels surrounded by the complex interfering features of in vivo skeletal muscle. Graded diameter changes and vasomotion were induced with pharmacological agents. Data obtained by auto-tracking agreed with data obtained by traditional manual methods of measurement. These findings indicate that a pattern recognition algorithm based on cross-correlation analysis provides a highly accurate method for real-time automatic tracking of microvessel diameters.
