Light sheet fluorescence microscopy images deblurring with background estimation

摘要

The deblurring of microscopy images has received a lot of attention in past decades, especially with regularisation involving the Total Variation semi-norm. In this paper we propose the evaluation of 3D point spread function space-invariant restauration methods with background estimation for 3D light sheet fluorescence microscopy images. A 3D image model for synthetic static multicellular tumor spheroids is introduced, allowing a more realistic evaluation for the proposed methods, underlining better results with low SNR values. Encouraging results on real data of cancer cells cultured as multicellular tumor spheroids are also presented.