Integrated spatio-spectral method for efficiently suppressing honeycomb pattern artifact in imaging fiber bundle microscopy

摘要

In a fiber bundle, spots of varying intensity occur throughout the fiber structure, and the ensemble of these spots creates the honeycomb pattern image, called a pixelated image. The spatio-spectral method, an integrated suppression method for the honeycomb pattern in the spatial and Fourier (spectral) domains, has demonstrated efficient restoration of microscopic images based on the fiber bundle. The morphological processing in the spatial domain partially interpolated the gaps between the circular patterns of fiber cores, and the selective band-reject filter in the Fourier domain was applied to reduce the honeycomb-shaped artifact. The auto-corrective search for the global peaks in the localized region was employed to apply the notch filter to the exact position of the individual periodic component in the Fourier domain. We found that a closing preprocessing in the spatial domain, combined with a Gaussian notch Fourier domain filter, yields the optimum image quality when eliminating the honeycomb pattern.