SUPPOSe 3Dge: A Method for Super-Resolved Detection of Surfaces in Volumetric Fluorescence Microscopy

Abstract

We present SUPPOSe 3Dge, a method for the detection of surfaces in volumetric fluorescence microscopy images with resolution better than the diffraction limit. The method works by approximating an arbitrary object surface as a polyhedron formed by small triangular faces, which are thought as virtual point like sources of signal gradient. Then, the super resolution arises from solving a least square problem only on the positions of the sources, by comparing the measured image gradient with a synthetic one built from the approximated surface and the known point spread function of the instrument. We present the formulation of the method and its characterization using simulated measurements of two different types. We show that the method is able to identify with high precision and accuracy the sizes and shapes of objects smaller than the instrumental resolution. Using different quantifiers that measure the distance between the solution and the target object, we report reconstructions with an accuracy 100 times better than the axial resolution and 40 times the lateral resolution of the instrument.