Lifetime-Aware Single-Cell Segmentation for Fluorescence Lifetime Imaging Microscopy

Lifetime-Aware Single-Cell Segmentation for Fluorescence Lifetime Imaging Microscopy

Guangchen Wang, Trinity College Dublin.

06th of November 2025

Abstract: 

Functional imaging is a class of medical imaging techniques that enable measurement of metabolic and biochemical activities in living cells and tissues. Recent advances in imaging equipment, particularly Fluorescence Lifetime Imaging Microscopy (FLIM), have made it possible to record high-resolution temporal signals at every pixel. Single-cell analysis using FLIM is crucial for quantifying cellular heterogeneity and connecting molecular dynamics to phenotypes. However, current FLIM cell segmentation methods typically operate on intensity averaged images, disregarding the temporal information. We propose a segmentation framework that uses the full FLIM signal, exploiting both morphological and decay information through a U-Net augmented with spatial-temporal self-attention. We demonstrate the proposed method on FLIM data and discuss the potential to generalize it to other types of functional imaging.