Spatially pooling photon information enables photon-efficient quantitative imaging

Quantitative fluorescence imaging like lifetime imaging and hyperspectral imaging derive molecular contrast from photons spread through spatial pixels and time or spectrum channels. In the realm of few photons, traditional pixel-by-pixel analysis disregards the spatial relationships between neighboring pixels dictated by the microscope's point-spread function (PSF).

We present SPOOL (Spatially Pooled Optical Observation Likelihood), a training-free Poisson inverse framework that recovers spatially distributed info without prior knowledge of emitter positions, spatial support or component assignments. SPOOL utilizes the PSF along with temporal-decay or spectral-response dictionaries to jointly recover source-space amplitudes and quantitative contrast.

The precision gain for an isolated source is determined by the PSF width in detector pixels, scaling with optical sampling rather than the physical source of contrast. The model predicts that lifetime precision gains linearly with the number of pixels spanning the PSF's full width at half maximum, a scaling confirmed by Monte Carlo simulations.

At one photon per foreground pixel, SPOOL reduces lifetime dispersion sixfold in fluorescent-bead experiments and cuts the lifetime root-mean-square error from 1.19 to 0.45 ns in dual-labeled cells. This framework remains applicable to hyperspectral imaging, extracting spectral contrast from generic emission bands without needing prior fluorophore spectra.