Probabilistic tracking of virus particles in fluorescence microscopy image sequences

Abstract

Fluorescence time-lapse microscopy is a powerful technique for observing the spatial-temporal behavior of viruses. To quantitatively analyze the exhibited dynamical relationships, tracking of viruses over time is required. We developed probabilistic approaches based on particle filters for tracking multiple virus particles in time-lapse fluorescence microscopy images. We employed a mixture of particle filters as well as independent particle filters. For the latter, we have developed a penalization strategy to maintain the identity of the tracked objects in cases where objects are in close proximity. We have applied the approaches to synthetic images and quantified their performance. We have also successfully applied the approaches to real microscopy images of HIV-1 particles and have compared the tracking results with ground truth from manual tracking.