Nonhomogenous random fields are known to be well adapted to modeling a wide class of images. Their computational complexity generally causes their lack of appeal, we propose a more efficient model capable of capturing textures, shapes, as well as jumps typically encountered in infra-red images. The so-called Levy Random fields as we show, can indeed represent a very well adapted alternative for...

In this article we consider the Levy processes and the corresponding semi-group. We represent the generator of this semigroup in a convolution form. Using the obtained convolution form and the theory of integral equations we investigate the properties of a wide class of Levy processes (potential, quasi-potential, the probability of the Levy process remaining within the given domain, long time b...