A Study on Optimal Parameter Estimation of Bilateral Filter Using Distribution Distance

Abstract

Bilateral Filter has properties of high noise removal. However, the estimating of suitable parameter by using an input image is required to obtain a fine denoising image. A method of parameter estimation using distribution distance has proposed. The method is optimal parameter estimation of bilateral filter based on distribution distance and assumption noise distribution that is estimated by differences between input and output images. However, there is problem if input images include many small edges. In this paper, we propose estimation of parameter using noise distribution that was estimated from estimation method that less susceptible to the original signal. Moreover, we show experimental results using proposed method.