Section: New Results

Lifetime estimation of moving vesicles in FLIM microscopy

Participants : Charles Kervrann, Philippe Roudot.

Fluorescence lifetime imaging microscopy (FLIM) is a widely spread imaging technique for sensing fluorophore environment in a living biological sample (like pH, ions...). Fluorescence lifetime (i.e. the average time a fluorophore stays in excited state before relaxing to its ground state possibly emitting a photon) is particularly useful to detect the Förster resonance energy transfer (FRET) which quantifies spatial proximity between molecules. We have proposed a statistical framework that exploits the intensity model of the frequency-domain FLIM output to jointly estimate trajectories and lifetimes of tracked vesicles. The proposed tracker, inspired from template cross-correlation or gaussian fitting, combines lifetime estimation and robust M-estimation in a efficient and fast way. Estimation of movement and lifetime are decoupled and alternatively performed, while particle/spot detection is performed on the first frame (Figure 4 ). To improve the results on real image sequences depicting moving vesicles, the background (cytoplasmic auto-fluorescence) model parameters and the scale parameters involved in the M-estimation procedure are estimated in our approach.

Partner:: F. Waharte and J. Boulanger (UMR 144 CNRS PICT IBiSA Institut Curie)

Figure 4. Lifetime map after spot/particle motion compensation on a phase-modulated fluorescence image stack.