Click here to flash read.
Light-sheet fluorescence microscopy (LSFM) is used to capture volume images
of biological specimens. It offers high contrast deep inside densely
fluorescence labelled samples, fast acquisition speed and minimal harmful
effects on the sample. However, the resulting images often show strong stripe
artifacts originating from light-matter interactions. We propose a robust
variational method suitable for removing stripes which outperforms existing
methods and offers flexibility through two adjustable parameters. This tool is
widely applicable to improve visual quality as well as facilitate downstream
processing and analysis of images acquired on systems that do not provide
hardware-based destriping methods. An evaluation of methods is performed on
LSFM, FIB-SEM and remote sensing data, supplemented by synthetic LSFM images.
The latter is obtained by simulating the imaging process on virtual samples.
No creative common's license