Incorporating variance within binary flow for leukocyte tracking

Tracking leukocytes in vivo is vital in determination of the roles and significance of integrins and selectins within the inflammation process and the effectiveness of anti-inflammatory drugs. In this paper, we propose a region-based snake for segmenting objects in the presence of inhomogeneities. We then use this geometric snake to track leukocytes in intravital microscopy. The results show that the binary flow algorithm introduced here improves upon the traditional binary flow algorithm and the gradient vector flow algorithm both in terms of percentage of frames tracked and the root mean squared error.