Enhanced resolution and sectioning capability via point spread function estimation and boosted speckle illumination

We present a computational imaging method that simultaneously enhances optical sectioning and spatial resolution using speckle illumination fluorescence microscopy. The process involves acquiring a set of fluorescence image frames with independent random speckle pattern illuminations, boosting the individual speckle images by applying regularized deconvolution using the point spread function (PSF) directly estimated from the speckle data, and calculating the pixel-wise standard deviation without further deconvolution. The approach yields an improved sectioning capability, comparable to confocal microscopy, while recovering high-frequency information up to a factor of x2 beyond the classical diffraction limit of confocal microscopy and matching the performance of 3D structured illumination microscopy. The technique is applicable to both 2D and 3D imaging and remains relevant under complex optical aberrations thanks to the robust PSF estimation from the multiple speckle illumination fluorescence image frames. Simulation and experimental results confirm the method’s effectiveness using metrics in both the frequency (Fourier ring correlation, FRC) and spatial domain (two-point resolution).