Isotropic spectrum optimization for enhancing image fidelity in super-resolution structured illumination microscopy.

Abstract

Super-resolution structured illumination microscopy (SR-SIM) performs spectral expansion of high-frequency information encoded in stripe patterns. However, using a limited number of pattern orientations (typically three) results in a petal-like frequency spectrum, leading to structural and intensity fidelity degradation in reconstructed images. In this Letter, we propose an integrated spatial-frequency domain SIM reconstruction method that enables isotropic spectrum expansion, called ISO-SIM. ISO-SIM overcomes structural artifacts caused by an anisotropic spectral expansion in traditional SIM imaging. We demonstrate the feasibility and fidelity of ISO-SIM through simulations, Argolight slide, and live-cell imaging. ISO-SIM enhanced structural similarity and reduced the error in the mean intensity ratio at certain spatial frequencies compared to Wiener-SIM. We further applied ISO-SIM to live-cell quantitative FRET imaging. ISO-SIM-FRET ensured that the measured FRET efficiency matched the ground truth, with a 19% reduction in the standard deviation compared to Wiener-SIM-FRET, maintaining intensity fidelity and enhancing the accuracy of quantitative analysis while suppressing artifacts.