An Overview of Denoising Methods for Different Types of Noises Present on Graphic Images

Abstract

It has been established that one of the reasons for the complication of the decision-making process is the deterioration of the quality of the input information obtained on the basis of various images due to overlaying noise on them, which may have different origin and characteristics. Studying a certain class of noise in the context of considering it as a function allows you to focus on determining its parameters, the degree of influence of these parameters and the artificial noise generation. An overview of the noise of different types and their effects was performed for further evaluation of the quality of recognition systems. Noises that arise in this case, are subject to classification in order to study, formalize and further eliminate or minimize their harmful effects. Studying a certain class of noise in the context of considering it as a function allows you to focus on determining its parameters, the degree of influence of these parameters and the artificial noise generation. Research shows that there are many types of noise that negatively affect the processing and analysis of images. An overview of various types of noise - Gaussian noise, shot noise (Poisson noise), noise type “salt and pepper” (impulse noise), noise of grains of a film, speckle noise, noise, giving a blur effect (they can be imposed with different degree of transparency); the features of overlaying such noise are determined. Also, the listed types of noise can be superimposed on each other. The method of logical generalization, overlaying of image noise using the Matlab environment is used. Comparison of several noise that creates the effect of blurriness when applied to images with varying degrees of transparency. Generating different noise leads to further overlay on real images of special noise masks with given parameters values - such as the intensity and size of the noise, the law of distribution of their centers, and so on.