Automated leukocyte detection in VIVO

In this paper we propose a technique for detecting rolling leukocytes within intravital microscopy based on Bayesian classification. The classification depends on a feature score, the gradient inverse coefficient of variation (GICOV), which serves to discriminate rolling leukocytes from a cluttered environment. An ellipse matching algorithm and B-spline snake for the aim of boundary extraction are discussed. The extracted contours that have a GICOV score above the analytically determined threshold are retained as identified objects. Experimental results using 327 rolling leukocytes were compared to those of human experts and currently used methods.