Additive noise generally refers to thermal noise, shot noise, etc., and their relationship is summed signal, whether or not a signal, such noise is always present. Usually the communications additive random noise of the system as a background.
Multiplicative noise is generally caused by the channel is not ideal, and their relationship is multiplied signal, a signal in which he will not, signals not. Multiplicative random as time-varying system (e.g., fading or Doppler) or caused by non-linear.
Multiplicative noise common to images of real-world applications of them, such as synthetic aperture radar, ultrasound, laser systems and other coherent image of them. With a standard additive white gaussian noise distinguished, multiplicative noise or gamma Rayleigh distribution function. Multiplicative serious noise pollution to the image, effectively multiplicative noise image processing more difficult, since more severe fluctuations multiplicative noise, low uniformity.
https://blog.csdn.net/sousky/article/details/89350093
It gives the relevant multiplicative noise process described in several ways.
Based on the multiplicative noise removing method homomorphic Wavelet
Noise analysis and medical ultrasound image denoising:
Medical ultrasound image denoising algorithms can be divided into two categories, one is performed using a denoising algorithm speckle noise distribution, another common method is based on some image processing algorithms, if wavelet-based medical ultrasound image to noise algorithm, based on anisotropic filtering method, based on non-neighborhood average algorithm.
Summary of conventional medical ultrasonic image denoising process from the spatial domain or in the transform domain, can be divided into two categories: one is the spatial domain denoising algorithm, comprising local adaptive algorithm, anisotropic diffusion filtering algorithm, non-neighborhood means algorithm, other hybrid algorithm. Another noise reduction filtering algorithm is the transform domain, including wavelet transform algorithm, the outline wave denoising algorithm and Fourier transforms.
