Super-resolution localization microscopy (SRLM) breaks the diffraction limit, making possible the observation ofsub-cellular structures. Challenges remain in SRLM due to a long data acquisition time. To overcome the limitation, themethods based on compressed sensing (CS) have been proposed. However, at the current stage, the widely usedsparsity-based localization methods, e.g., interior point method (IPM), is computationally intensive. To address theproblem, in this paper, we introduce an alternative CS reconstruction method to super-resolution imaging model, whichis achieved by using gOMP (generalized Orthogonal Matching Pursuit). A series of numerical simulations with varyingemitter densities and signal-to-noise rations (SNRs) are performed to evaluate the performance of gOMP method. Theresults show that whatever gOMP or IPM is used in SRLM, the obtained localization accuracy is similar. But, thedata-processing time of gOMP can be significantly reduced (> 100 times) than the previous reported IPM method. As aresult, gOMP provides the potential for reducing the computational cost while maintaining a desired spatial resolution,which is beneficial for SRLM imaging.
展开▼
