Biomedical Application of Target Tracking in Clutter

摘要

The movement of leukocytes (white blood cells) and their interaction with the endothelium (vessel wall) provides valuable information about the mechanism of inflammation and inflammatory disease. To investigate leukocyte motion within living animals, advanced automated tracking algorithms are requisite. The authors introduce military target tracking algorithms for use in tracking cell movement. Their dataset consists of intravital microscopy video recordings of rolling leukocytes in the mouse cremaster muscle observed via transillumination. They tracked 33 cells from 10 venules (small vessels). Five sets were TNF-alpha treated venules and five sets were untreated. The TNF-alpha treatment increases the inflammatory response and thus slows down the rolling cells. In 33 experiments, they compared the performance of five trackers that utilize video microscopy technology. The trackers tested include the centroid tracker, the correlation tracker, an enhanced centroid tracker, an enhanced correlation tracker, and an active contour (snake) tracker. Of the five methods, the snake tracker proved to be the most robust in terms of the highest percentage of frames tracked and the lowest root mean-squared error. This paper provides an overview of the five trackers and results from the 33 experiments. (1 table, 4 figures, 3 refs.).