Image noise detection and identification using abstract matrix and difference matrix techniques

Abstract EN

The problem of noise identification in one-dimensional (signal) and two dimensional (image) processing is very important in engineering and computer science applications.

These noises: salt, pepper, Gaussian and speckle, are different in subjective and objective behaviors.

In this paper we firstly check the observed image if it is noisy or noise-free (original) image.

Then we propose methods for identification the noise in the observed image named the Abstract Matrix (RM) and the Difference Matrix (DM) methods.

AM and DM are matrices that generated from the gray levels of the image.

We generate the RM and the DM of the image in the spatial domain; their generations are very fast.

According to the distribution of the elements of the AM and DM, we can predict first whether the observed image is the original (it does not need denoising process) or the distorted version (it does need a denoising process) and second identify the noise types.