Huang, X. et al. Nat. Biotechnol. 36, 451–459 (2018).

Structured illumination microscopy (SIM) offers twice the resolving power of diffraction-limited microscopy with relatively low doses of light compared with those required for other super-resolution modes, which makes it useful for live imaging. However, SIM involves image reconstruction that is prone to artifacts. Huang et al. have developed an approach called Hessian-SIM to reduce artifacts in SIM image reconstruction for improved fast live-cell super-resolution imaging. Hessian-SIM uses a deconvolution algorithm based on Hessian matrices that makes use of a priori knowledge of an imaged structure to guide image reconstruction. This deconvolution algorithm outperforms current algorithms at low signal intensities and allows imaging at a fraction of the photon dose of conventional SIM. Using their approach, the researchers carried out hour-long time-lapse imaging of actin filaments in live cells and were able to image never-before-observed structures involved in endocytosis.