Mathematical Models of Images and Video and Its Application

In this paper, we propose mathematical models of images and video. The models are based on the approximation of images and video by multidimensional discrete Markov random processes with any number of states. To validate both models, the real and artificial images with the same the transition probability matrices are given. To investigate the adequacy of the models, the estimation of the transition probability matrices was evaluated. According to the experimental results, the presented models are adequate to real images and video, and have high computational efficiency. As an application, the method of the restoration of partially distorted image is then provided. Experimental results show the effectiveness of our proposal.