Super-resolved edge detection in optical microscopy images by superposition of virtual point sources

Abstract

A new approach to the edge detection problem is presented which is specially designed to achieve high accuracy detection, below instrumental resolution (super resolution) in microscopy images. The method is based in a modified version of a recently published algorithm known as SUPPOSe, which performs a numerical reconstruction of an image as a superposition of virtual point sources. The method was tested in simulated and experimental optical microscopy images and compared to the standard Laplacian of Gaussian algorithm, showing huge differences when the size of the object is smaller than the lateral resolution provided by the instrument.