A new objective function for super-resolution deconvolution of microscopy images by means of a genetic algorithm

LACAPMESURE, AXEL; MARTINEZ, SANDRA; O.E. MARTÍNEZ

GECCO 2020 Companion - Proceedings of the 2020 Genetic and Evolutionary Computation Conference Companion

Association for Computing Machinery, Incacmhelp@acm.org

Año: 2020 p. 271 - 272

The SUPPOSe algorithm for super-resolution image deconvolution relies in assuming that the image source distribution can be modeled as a superposition of point sources of equal intensities, achieving a fivefold improvement in the spatial resolution. A genetic algorithm is used to find the positions of the sources by optimizing an objective function that also depends on their intensities. In this work we present a new objective function for the SUPPOSe algorithm that is independent of the source intensities. We compare both methods and prove the same performance but without the necessity to fit the intensities. This will allow to replace the optimization of the intensities with an optimization step to fit the instrument response function for a blind deconvolution.