Fig. 2: Systematic validations using SMLM simulations.

a Simulations of 2D-SMLM with homogeneous illumination (inducing overall spatial invariant active density), with high-density (HD, top) and low-density (LD, bottom) emitting fluorophores in each frame. Cyan and white arrows represent the errors found by the RSM or rFRC map, respectively. b 2D-SMLM simulation with inhomogeneous illumination (high intensity in the center and decreasing illumination toward the edges). c 2D-SMLM simulation with inhomogeneous focus (in focus at the center and defocusing toward the edges). From left to right: Merged MLE reconstructions (green channel) and ground-truth images (red channel); Spatial subtractions between ground-truth images and MLE reconstructions; Spatial subtractions between wide-field ground-truth images and wide-field images generated from the MLE reconstructions, a.k.a. the RSM; The full SQUIRREL map (RSM corresponding to the green channel and FRC map to the magenta channel); The rFRC maps of two MLE reconstructions from odd frames and even frames, respectively; The full PANEL visualizations (RSM corresponding to the red channel and rFRC map to the green channel). Scale bars: (a) 500 nm; (b, c) 1 μm