Improvement in quantitative analysis of RBCs velocity in microcirculation based on block-matching motion estimation

Abstract

This paper presents an improved method towards a semi-automated method for the quantification of red blood cells (RBCs) velocity in individual microvessel. Our aim is to use straightened vessel centerline in individual microvessel, which offers a main benefit to reduce computational time for quantitative analysis of blood flow velocity based on block-match motion estimation. The quantitative method used the modified block-matching method based on straightened vessel centerline image to perform a wide range of changes in amount of search comparison, subsequently, to use the optical flow method for fine adjustment pixel by pixel and to complete the overall velocity estimation. In the evaluation experiment, the current method and the proposed one were applied to make tests on simulated vessel images for performance comparisons. The estimation results are quite rapid and accurate.