基于压缩感知STORM超分辨成像与像素的关系

Relationship of STORM super-resolution imaging and pixel based on compressive sensing

针对基于压缩感知STORM(stochastic optical reconstruction microscopy)超分辨成像效果差的不足,提出采用高分辨相机改善基于PSF测量矩阵性能的方法。该方法能够改善基于PSF测量矩阵的约束等距性(restricted isometry property,RIP),从而达到提高重构效果的目的。实验结果表明,采用高分辨相机后基于PSF(point spread function)测量矩阵的列不相关性更好,重构能力、定位准确度和识别率都得到极大改善。同时探讨了以传统指标体系评价基于压缩感知的超分辨重构质量的优劣和适用性。发现匈牙利法和质心法的组合方案较能反应真实的基于压缩感知的超分辨重构效果。

In view of the deficiency of the STORM （stochastic optical reconstruction microscopy） super resolution imaging based on compressive sensing, a method is proposed to improve the performance of PSF measurement matrix by a higher resolution camera. This method can improve the RIP （ restricted isometry property） of measurement matrix based on PSF （ point spread function） , so as to achieve the goal of improving the reconstruction effect. The experimental results show that column incoherence of measurement matrix based on PSF is better. Reconstruction ability, localization accuracy and detection rate have been greatly improved after the high resolution camera is used. At the same time, the advantages and disadvantages of the traditional indicator system to evaluate the quality of super resolution reconstruction based on compressive sensing are discussed. It is found that the combination of the Hungarian method and the centroid method can better show the effect of the super resolution reconstruction based on compressive sensing.