Second-order optimized regularized structured illumination microscopy (sorSIM) for high-quality and rapid super resolution image reconstruction with low signal level

Abstract

Structured illumination microscopy (SIM) is a widely used super resolution imaging technique that can down-modulate a sample's high-frequency information into objective recordable frequencies to enhance the resolution below the diffraction limit. However, classical SIM image reconstruction methods often generate poor results under low illumination conditions, which are required for reducing photobleaching and phototoxicity in cell imaging experiments. Although denoising methods or auxiliary items improved SIM image reconstruction in low signal level situations, they still suffer from decreased reconstruction quality and significant background artifacts, inevitably limiting their practical applications. In order to improve the reconstruction quality, second-order optimized regularized SIM (sorSIM) is designed specifically for image reconstruction in low signal level situations. In sorSIM, a second-order regularization term is introduced to suppress noise effect, and the penalty factor in this term is selected to optimize the resolution enhancement and noise resistance. Compared to classical SIM image reconstruction algorithms as well as to those previously used in low illumination cases, the proposed sorSIM provides images with enhanced resolution and fewer background artifacts. Therefore, sorSIM can be a potential tool for high-quality and rapid super resolution imaging, especially for low signal images.