A NOVEL ALGORITHM BASED ON CELLULAR AUTOMATA TO ELIMINATE NOISE IN DIGITAL IMAGES

Abstract

A cellular automaton is a processing system that makes decisions based on the information of its neighbors. Implementing a cellular automaton can have a direct relationship with the form of digital images representation, whereby it is possible to perform image processing applications using this concept. Due to this, this document presents the proposal of an algorithm based on cellular automata for the elimination of impulsive noise in digital images, which makes use of the adaptation mechanism to adjust to the environment conditions (in this case, the image) at the moment that the information acquired by the cellular automaton is insufficient to make a decision about the pixel under evaluation. In addition, thanks to the consideration of related works, an evaluation method can be established to observe the performance of the proposed algorithm against two algorithms presented by other authors. Tests are carried out with four images that have different characteristics. In evaluating the images exposed to different noise levels, the results obtained show that the proposed algorithm presents better according to the Structural Similarity Index (SSIM), since at noise levels between 10% and 90% improvements in noise reduction range between 15% and 68%.