Quasi-Periodic Image Noise Filtering Using Inverse Laplace Filter

Abstract

In the process of image acquisition a lot of different types of noise can arise. Filtering this noise is a well-known problem in image processing. Quasi-periodic noise filtering has proven to be a challenging problem in this topic. This paper presents a novel automated approach for filtering quasi-periodic noise using Laplace distribution for modelling Fourier transform coefficients, Inverse Laplace filter and Median filter. Results show that this approach can filter the noise even when the noise is more present in the image than the original content of the image. Average increase in the amount of the signal of the image is 50%.