Adaptive Frequency Domain Optimization for Artifacts‐Free Structured Illumination Microscopy

Structured illumination microscopy (SIM) has emerged as a key tool for live‐cell super‐resolution imaging due to its low phototoxicity and high speed. However, its widefield nature makes it susceptible to out‐of‐focus background, which introduces stripe‐like or honeycomb artifacts, that compromise image fidelity. Existing solutions have significant limitations: polarization‐modulated SIM requires additional complex optics; post‐processing algorithms often rely on subjective parameter tuning and can sacrifice high‐frequency information; while preprocessing methods based on spatial domain statistics or fixed physical models risk signal over‐subtraction or inaccurate estimation. In this work, we present a frequency‐domain optimization method based on the established relationship between out‐of‐focus signal and frequency‐domain linearity to eliminate the artifacts caused by background interference. By designing an adaptive bandpass filter and applying it to raw data, the low‐frequency background component is effectively separated prior to reconstruction. This preprocessing step eliminates artifacts and ensures a stable, uniform signal‐to‐background ratio under diverse imaging conditions, while preserving the theoretical resolution of SIM. Being computationally efficient and compatible with various SIM modalities, our method offers a robust and practical solution for enhancing the fidelity of nanoscale biological imaging.