Slicing Network for Wide-Field Fluorescence Image Based on the Improved U-Net Model.

Abstract

Fluorescence imaging stands as a pivotal component in biomedical research, requiring the elimination of out-of-focus background noise resulting from wide-field volumetric illumination of the whole field-of-view and scattering within thick biological tissues. Traditional methods struggle to effectively address varying degrees of defocusing in fluorescence images. This study introduces the utilization of upU-Net, 3D U-Net, and 3D upU-Net as defocusing networks tailored for 2D and 3D wide-field fluorescence images, yielding notable enhancements. These advancements facilitate more economically viable confocal microscopy, delivering significant advantages to biologists presently utilizing wide-field fluorescence microscopy.