A noised model and its application for detecting image materials

Abstract

Images always have noises due to image acquisition processes. Noises are created by self of objects and environments. Each of materials has specific number of noises, which characterizes that material. Noise features of the material are called material noises. This paper deals with a technique for detecting image materials which base on noise analysis of images. By experiments, the technique exposed  fairly accurate results for some natural materials at first, especially homogeneous materials about illuminates and extracted invariant features is sparse.