In fluorescence microscopy, Single Molecule Localization Microscopy (SMLM) techniques aim at localizing with high precision high density fluorescent molecules by stochastically activating and imaging small subsets of blinking emitters. Super Resolution (SR) plays an important role in this field since it allows to go beyond the intrinsic light diffraction limit.
In this work, we propose a deep learning-based algorithm for precise molecule localization of high density frames acquired by SMLM techniques whose L2-based loss function is regularized by positivity and L0-based constraints. The L0 is relaxed through its Continuous Exact L0 (CEL0) counterpart.The arising approach, named DeepCEL0, is parameter-free, more flexible, faster and provides more precise molecule localization maps if compared to the other state-of-the-art methods. We validate our approach on both simulated and real fluorescence microscopy data.
The code of this paper is built upon the ZeroCostDL4Mic and Deep-STORM repositories.
Please consider to cite DeepCEL0 if you find it helpful.
@article{cascarano2022deepcel0,
title={DeepCEL0 for 2D single-molecule localization in fluorescence microscopy},
author={Cascarano, Pasquale and Comes, Maria Colomba and Sebastiani, Andrea and Mencattini, Arianna and Loli Piccolomini, Elena and Martinelli, Eugenio},
journal={Bioinformatics},
volume={38},
number={5},
pages={1411--1419},
year={2022},
publisher={Oxford University Press}
}