The characterization of scanning noise and quantization on texture feature analysis

The study of the effects of scanning on texture features is of great interest to computer-based screening systems. A mathematical model is developed for understanding how the original image gets distorted due to the contrast variability and geometric distortion inherent in the scanning process. Both quantitative and qualitative results (for sixty common texture features) are given.