| [ KAIST ] in KIDS 글 쓴 이(By): guest (Grizzly) <cortex.stat.wisc> 날 짜 (Date): 2001년 9월 8일 토요일 오전 08시 17분 52초 제 목(Title): Re: [질문] Gaussian filter Gaussian filter is usually refered as a Gaussian kernel smoothing and used to increase the signal-to-noise ratio (SNR). For a given n-dimensional signal f(x), it is defined as the convolution of f(x) with the Gaussian kernel K(x), which is the standard normal distrbution. There are three ways of implementing the numerical algorithms for this filter. 1) By applying discrete Fourier transform followed by the inverse Fourier transform 2) Discretizing the Gaussian kernel smoothing as a linear summation 3) Since the Gaussian filter is the solution of an isotropic diffusion equation, by actually solving the diffusion equation. Each approach has its own merits and can be easily implemented in matlab. |