Parameter Customization of Bilateral Filtering Image
Wenjia Ding, Yi Xie, Yulin Wang · 2018
This paper is to improve the bilateral filtering algorithm by customizing the parameters so as to denoise the blurred image edges better and quicker. By the analysis of bilateral filtering algorithm, we find that two important parameters affecting filtering performance are Gaussian standard spatial and pixel domain deviation. In this paper, we aim to analyze and compare variable relationship between parameters and image noise variance. The derived results are verified by different set of experiments, such as low- and high-frequency image, grayscale and color image, and images of different sizes and images with different subjects. The obtained results can be used to reduce the time for the selection of each parameter in image pre-processing applications, such as traffic scene image and medical image denoising. Further experiments are carried out to apply it to the actual being.