A New Method for Generating Quasi-Periodic Noise on Digital Images

A novel approach for generating quasi-periodic noise, a pattern-like interference that is quite common in medicine and satellite imaging, is presented in this paper. Method consists of simple and intuitive spatial-domain image processing techniques that outperform the similar approaches in terms of diversity. Our algorithm is highly configurable with very intuitive tuning parameters. The key improvement of our generator is the introduction of new parameters: amplitude, frequency and slope of the sinusoidal waves. Besides the algorithm, two new metrics are proposed in this paper for relative comparison of generators - deviation and diverseness. As these metrics suggest, the noise generator presented in this paper shows a high level of quasi-periodic noise diversity. Improvement in diversity is achieved through three new parameters proposed in comparison to previous research done in this field of generating noise.