Dark Sectioning Clears Blurry Backgrounds in Fluorescence Microscopy Images

What is it about?

When scientists look at cells under a fluorescence microscope, they often see a "fog" or blurry background that hides the real structures they want to study. This happens because the microscope captures light not just from the focused layer, but also from above and below it. Traditionally, researchers had to buy expensive hardware upgrades (like confocal or light-sheet microscopes) or use complicated physical methods to remove this fog. Our team developed a new software method called Dark Sectioning that can automatically remove this blurry background from fluorescence microscope images. The idea came surprisingly from a technique used in outdoor photography to remove haze from landscape photos. We adapted this "dark channel prior" concept for microscopy by combining it with frequency-domain separation and knowledge about how microscopes focus light. Dark Sectioning works with almost any type of fluorescence microscope—including wide-field, confocal, two-photon, light-sheet, STED super-resolution, and structured illumination microscopy (SIM). It requires only a single image (no need for multiple exposures or special hardware) and can be applied as a post-processing step to existing data. In tests, it improved image clarity by nearly 10-fold and made it much easier for computers to automatically identify and count cells or organelles in thick tissue samples—such as neurons in mouse brain tissue, nuclei in prostate cancer sections, and mitochondria in living cells. We have made the software freely available as open-source code (Fiji plugin, MATLAB, and Windows executable) so that any researcher can use it without purchasing new equipment.

Featured Image

Photo by National Cancer Institute on Unsplash

Photo by National Cancer Institute on Unsplash

Why is it important?

Removing out-of-focus background has been a persistent challenge in fluorescence microscopy for decades. Existing solutions force researchers into difficult trade-offs: expensive hardware upgrades, reduced imaging speed, photobleaching of delicate samples, or limited compatibility with specific microscope types. Dark Sectioning is unique because it is universal, hardware-free, and high-fidelity. Unlike denoising or deconvolution methods that address different problems, Dark Sectioning specifically targets and removes out-of-focus background while preserving weak in-focus signals. It outperforms commercial solutions (such as Leica's ICC algorithm) and open-source alternatives (Rolling Ball, Sliding Paraboloid, Sparse deconvolution) in both visual quality and quantitative metrics. The timeliness of this work is significant: as biological imaging moves toward high-content screening, whole-slide imaging, and AI-driven analysis, the quality of input images becomes critical. Dark Sectioning can immediately improve the accuracy of automated segmentation and analysis pipelines without requiring any hardware investment. It also enhances advanced techniques like SIM, SOFI, and deep-learning super-resolution by providing cleaner raw data for reconstruction. In practical terms, this means a researcher with a basic wide-field microscope can now achieve optical sectioning quality comparable to much more expensive confocal systems—simply by processing their images with our free software. This democratizes high-quality imaging and could accelerate discoveries in cell biology, pathology, and drug screening.

Perspectives