Microscopy creates images of objects which should represent the nature of the object as well as possible. Fluorescent light, which emanates from the object, passes through the various optical elements of the beampath and eventually gets collected by the detector. Unfortunately, on the way to the detector the signal is changed in such a way, that the quality of the resulting image suffers. As a consequence, the image is never a 100% correct representation of the object. This effect is strongest seen in a classical widefield fluorescence microscopy which does in fact not offer any optical sectioning capability, but also exists to a different degree in optical sectioning microscope systems, e.g. Confocal, Lightsheet or ApoTome.

Fortunately, the dominant function which has this deleterious effect on the image, is based on the optical design principles of the light microscope and therefore well understood. We call this the point spread function (PSF) of the microscope system.

Deconvolution is a mathematical method which can reverse the effect of the PSF on the image and can therefore to a large extent restore the image to better represent the object. In the case of widefield imaging, Deconvolution can even convey optical sectioning properties to the result image allowing true three-dimensional restoration.

The following improvements can be obtained by using deconvolution:

• Denoising
• Removal of out of focus light -> deblurring, improved contrast
• Increasing signal to noise ratio by reassigning photons
• Restoration of sparsely sampled data
• Increase of resolution in X, Y and Z

As the object and the way, it was prepared, becomes part of the optical system during imaging, the largest variable to consider when doing deconvolution is the sample itself. Since the sample conditions can vary widely, information about the sample needs to be provided to the deconvolution function. The better the sample conditions are known, the better the outcome will be.