Applying Space-Variant Point Spread Function to Three-Dimensional Reconstruction of Fluorescence Microscopic Images

Abstract

Three-dimensional (3D) reconstruction of fluorescence microscopic images is a challenging topic in the image processing, because the imaging system is very complex, and the point spread function (PSF) continuously varies along the optical axis. Generally, the more exact the PSF is, the higher the reconstruction accuracy is. An image reconstruction method is proposed for fluorescence microscopic sample based on space-variant PSF (SV-PSF) which is generated by cubic spline theory in this paper. Firstly, key PSFs are estimated by blind deconvolution algorithm at several depths of fluorescence microscopic image stack along the optical axis. Then, other PSFs are interpolated using cubic spline theory. Finally, a 3D microscopic specimen model is reconstructed by this group of SV-PSFs. The experimental results show that the proposed method is obviously superior to the method in which space-invariant (SI) PSF is used to reconstruct the simulated and real fluorescence microscopic images.