Liu Sheng, Kromann Emil B, Krueger Wesley D, Bewersdorf Joerg, Lidke Keith A
Opt Express. 2013 Dec 2;21(24):29462-87. doi: 10.1364/OE.21.029462.
Localization-based superresolution imaging is dependent on finding the positions of individual fluorophores in a sample by fitting the observed single-molecule intensity pattern to the microscope point spread function (PSF). For three-dimensional imaging, system-specific aberrations of the optical system can lead to inaccurate localizations when the PSF model does not account for these aberrations. Here we describe the use of phase-retrieved pupil functions to generate a more accurate PSF and therefore more accurate 3D localizations. The complex-valued pupil function contains information about the system-specific aberrations and can thus be used to generate the PSF for arbitrary defocus. Further, it can be modified to include depth dependent aberrations. We describe the phase retrieval process, the method for including depth dependent aberrations, and a fast fitting algorithm using graphics processing units. The superior localization accuracy of the pupil function generated PSF is demonstrated with dual focal plane 3D superresolution imaging of biological structures.
基于定位的超分辨率成像依赖于通过将观察到的单分子强度模式拟合到显微镜点扩散函数（PSF）来找到样品中单个荧光团的位置。对于三维成像，当PSF模型没有考虑光学系统的特定系统像差时，光学系统的特定系统像差会导致定位不准确。在这里，我们描述了使用相位恢复的光瞳函数来生成更准确的PSF，从而实现更准确的三维定位。复值光瞳函数包含有关特定系统像差的信息，因此可用于生成任意离焦的PSF。此外，它可以被修改以包括深度相关的像差。我们描述了相位恢复过程、包含深度相关像差的方法以及使用图形处理单元的快速拟合算法。通过对生物结构的双焦平面三维超分辨率成像，证明了光瞳函数生成的PSF具有更高的定位精度。
