Parametric Variations of Anisotropic Diffusion and Gaussian High-Pass Filter for NIR Image Preprocessing in Vein Identification

Abstract

Near infrared (NIR) imaging is one of the promising methods for identification of superficial veins and widely researched and used in clinical medicine and biomedical studies. However, just like imaging in visible spectrum, NIR imaging is not adequate for exact recognition of the vein system as it is, therefore nearly every research starts with preprocessing to prepare the images for identification. Two major filtering methods are anisotropic diffusion and Gaussian high-pass filter which both consist of mandatory parametric adjustments for better visualization of the images and for revealing the vein system. Therefore in this paper we deal with parametric variations of these two methods on a NIR image to give ideas for choosing proper preprocessing techniques and parameters, excluding edge detection and vein detection methodologies.