Optimal Multivariate Gaussian Fitting for PSF Modeling in Two-Photon Microscopy

Tim Tsz-Kit (1) , Emilie Chouzenoux (2, 3) , Claire Lefort (4) , Jean-Christophe Pesquet (3)

1 Department of Mathematics [Hong Kong]
2 LIGM - Laboratoire d'Informatique Gaspard-Monge
3 GALEN-POST - Organ Modeling through Extraction, Representation and Understanding of Medical Image Content
4 XLIM-PHOT - Photonique Fibre et Sources Cohérentes

Fitting multivariate Gaussian functions constitutes a fundamental task in many scientific fields. However, most of the existing approaches for performing such fitting are restricted to 2 dimensions and they cannot be easily extended to higher dimensions. One of the main applicative areas where it is necessary to go beyond the existing techniques is the modeling of Point Spread Functions in 3D imaging. In this paper, a novel variational approach is proposed to fit multivariate Gaus-sians from noisy data in arbitrary dimensions. The proposed FIGARO algorithm is applied to two-photon fluorescence microscopy where its excellent performance is demonstrated.

Mots clés: optimization; proximal methods; PSF identification; two-photon microscopy; image restoration

Domaines: Traitement du signal et de l'image [eess.SP]

Tim Tsz-Kit, Emilie Chouzenoux, Claire Lefort, Jean-Christophe Pesquet. Optimal Multivariate Gaussian Fitting for PSF Modeling in Two-Photon Microscopy. 15 th IEEE International Symposium on Biomedical Imaging (ISBI 2018), Apr 2018, Washington DC, United States. ⟨10.1109/ISBI.2018.8363621⟩. ⟨hal-01803413⟩