Particle Tracking in Microscopy Images

             Leila Muresan 

      (Johannes Kepler Univ. Linz):


Abstract:

A frequent task in microscopy image processing is tracking of
undistinguishable spots in a sequence of images. The problem can be divided
into two parts: spot detection and data association, the former strongly
influencing the outcome of the latter. The spot detection phase is followed
by a fitting process, in order to achieve sub-pixel resolution. The low
signal to noise ratio of the images may result in poor quality of detection
and fitting, respectively. The data association phase is implemented based
on three methods: linear assignment problem, greedy matching and the IPAN
method. The IPAN method proved to be the most robust to missed spots and
overdetection. The necessity of establishing a quality measure for the
detected trajectories, as well as for the tracking process itself, is also
discussed.