Structured illumination microscopy (SIM) is a widely available super-resolution technique for bioscience, especially for living cell research, due to its high photon efficiency. However, the quality of SIM depends extremely on the post-processing algorithms (parameter estimation and image reconstruction), where parameter estimation is the critical guarantee for successful super-resolution reconstruction. In this letter, we present a novel SIM approach based on principal component analysis (PCA-SIM) that statistically purifies experimental parameters from noise contamination to achieve high-definition super-resolution reconstruction. Experiments demonstrate that our method achieves more accurate (0.01 pixel wave vector and 0.1% of 2π initial phase) parameter estimation and superior noise immunity with an order of magnitude higher efficiency than conventional cross-correlation-based methods, offering the possibility of faster, less photon dose, longer duration living cell SIM.
|