Super-resolution reconstruction of structured illumination microscopy based on pixel reassignment

Abstract

In this work, we report a pixel reassignment based super-resolution reconstruction algorithm for structured illumination microscopy (entitled PR-SIM). PR-SIM provides a twofold theoretical resolution enhancement by reassigning the pixels in raw SIM images with respect to the center of each illumination fringe and applying further deconvolution. By comparing with frequency domain based algorithms, PR-SIM is more immune to fringe distortion and, hence, it is more suited for large-field SIM in that it processes the raw images locally. Meanwhile, the reconstruction speed of PR-SIM can be enhanced by skipping empty regions in the image and further enhanced by employing GPU-base parallel calculation. Overall, we can envisage that the PR-SIM can be extended for other illumination modulation based microscopic techniques.