Richard-Lucy deconvolution algorithm for line-scanning confocal microscope images

Abstract

A Confocal laser scanning microscopy has been extensively used in biological researches. However, its insufficient scan speed or frame rate cause to some limits for in-vivo imaging. A line-scanning confocal microscope is an effective way to compensate the defect. A line-scanning confocal microscope, a variant of a confocal microscope, uses line-scanning in one direction instead of point-scanning. This technique increases scan speed significantly, but it also causes image resolution deterioration slightly. To overcome resolution deterioration, this study proposes an image deconvolution algorithm, based on the Richardson-Lucy algorithm with Total Variation regularization, assuming that there are two observed line-scanning confocal images whose scanning directions are perpendicular with each other. This line-scanning model is practical because of significantly faster scan time of a line-scanning confocal microscope than that of a point-scanning one. We will provide the validity of our algorithm by performing numerical experiments on 2D synthetic images.