Abstract
The SUPPOSe enhanced deconvolution algorithm relies in assuming that the image source can be described by an incoherent superposition of virtual point sources of equal intensity and finding the number and position of such virtual sources. In this work we describe the recent advances in the implementation of the method to gain resolution and remove artifacts due to the presence of fluorescent molecules close enough to the image frame boundary. The method was modified removing the invariant used before given by the product of the flux of the virtual sources times the number of virtual sources, and replacing it by a new invariant given by the total flux within the frame, thus allowing the location of virtual sources outside the frame but contributing to the signal inside the frame.
Highlights
The challenge of obtaining high-resolution images beyond the limit imposed by diffraction, aberrations and noise with a single image was pursued for decades by deconvolution of the image with the instrument response function or point spread function (PSF)
First we reconstruct two synthesized images using tiles by the original SUPPOSe and by the modified SUPPOSe method. In both synthetic images we observe that the modified SUPPOSe method removes the artifacts that appear in the intersection of the tiles when applying the original method. These artifacts are removed maintaining the same resolution that we had obtained when applying SUPPOSe to images supported inside the image frame
In this work we introduced a new fitting function for the method SUPPOSe presented in [12] to remove artifacts arising from the boundary
Summary
The challenge of obtaining high-resolution images beyond the limit imposed by diffraction (resolution beyond the Abbe limit), aberrations and noise with a single image was pursued for decades by deconvolution of the image with the instrument response function or point spread function (PSF). Because a background would be a dense signal an image with background is not adequate for methods assuming sparsity of the solution To solve this problem in [12] we needed to redefine the fitting function in the cases where there was an unknown background by subtracting the average value of the image and fitting with a PSF with null average. This was not taken into account in the original version of SUPPPOSe [12] where N was assumed equal to the sum of the signal within the frame To remove these artifacts in this paper we modified the optimization problem, changing the fitting function by a new one. We show that with the modified method we can remove artifacts of tiled F-actin wide-field images taken with high NA objectives In this case we could distinguish structures that are distorted when applying the standard SUPPOSe method
Sign-in/Register to view highlights & summary 
Published Version (
Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call