Enhancing the Visualization of the Microvasculature of Extrahepatic Bile Ducts Obtained from Confocal Microscopy Images

摘要

Confocal microscopy is an important tool for visualizing 3D datasets of fluorescent specimens, and has been used to investigate the structure of biological specimens. However, such images are affected by the noise introduced during the specimen preparation and image acquisition processes. Anisotropic diffusion is a non-linear filter that can significantly improve image quality while removing noise without blurring edges. This study investigates the application of anisotropic diffusion in confocal microscopy images by exploring different models for parameters' calculation. Our data consists of several slices of extrahepatic bile ducts containing a network of small vessels named Peribiliary Vascular Plexus (PVP), which are affected by several sources of noise. Experimental results show that anisotropic diffusion improved the volumetric visualization of the PVP. We validated the results using MSE and PSNR quantitative approaches and qualitative description by an expert user.