Fig. 1: Bayesian-SIM reconstruction with spatial domain noise propagation.

a SIM involves collecting fluorescence images (left) illuminated by structured intensity patterns (center). We utilize a point-spread function model (center)(PSF) and calibrate the camera noise parameters (right): gain (g), offset (o), and readout noise (σ). To infer the underlying fluorescence intensity map, we start from a small region (orange) and sweep over the entire sample in a parallelized fashion. b An image formation model, \({\mathcal{M}}\), involves the multiplication of an underlying fluorescence intensity map with illumination patterns and convolution with the PSF to obtain a noiseless image, μ, which is then corrupted by photon shot noise and camera noise. c A Monte Carlo algorithm to sample candidate fluorescence intensity maps from the posterior probability distribution. Based on the current sample (orange) we propose a new sample for the fluorescence intensity map (blue) and compute its corresponding posterior probability. Favoring higher probability, we stochastically accept or reject the proposed fluorescence intensity map and update the current fluorescence intensity map. We average many samples of fluorescence intensity maps after convergence to obtain B-SIM reconstruction (right). WF and B denote widefield image and B-SIM reconstruction, respectively.