Confocal-inspired computational fluorescence microscopy via transfer point spread function

Laser scanning confocal microscopy provides optical sectioning but is limited by slow acquisition and hardware complexity. Widefield fluorescence microscopy offers high-throughput, low-cost imaging, yet global illumination generates substantial out-of-focus and multiply scattered background. Here we present computational confocal fluorescence microscopy (CCFM), a physics-grounded framework that leverages a transfer point spread functon(TPSF) to reconstruct confocal-like volumes from widefield data. CCFM models widefield imaging as a cascaded process in which scattering-mediated redistribution is followed by system blur, enabling reconstruction beyond standard deconvolution. After a single calibration from paired widefield and confocal volumes, the TPSF can be reused for subsequent widefield imaging under matched imaging conditions and preprocessing strategy, without hardware modification. Across diverse specimens and imaging configurations, CCFM suppresses scattering background and restores optical sectioning beyond standard deconvolution, enabling confocal-like dynamic three-dimensional imaging at widefield speed.