In any image processing system denoising of images is an important step. The images can be corrupted by different noises with different levels. There are three types of noises available: impulse, Gaussian and Speckle noises with mixture of them. Many algorithms are proposed to remove salt & pepper (impulse) noise as well as Gaussian noise. The Robust statistics based filter is also proposed to remove either impulse or Gaussian noise using Lorentian rho function based robust M estimator. However, there is still a need to find a most efficient filter for image denoising, which can be effective for salt & pepper noise with different noise levels. In this paper we evaluate the performance of MM-estimator and M-estimator based image denoising filters for salt & pepper noise only. The results show very good impulse noise removal by MM estimator compared to M-estimator.
展开▼