Image Restoration with Mixed or Unknown Noises

This paper proposes a simple model for image restoration with mixed or unknown noises. It can handle image restoration without assuming any prior knowledge of the noise distribution. It is particularly useful for solving real life image restoration problems, since under various constraints, images are always degraded with mixed noise and it is impossible to determine what type of noise is involved. The proposed model can remove mixed type of noises as well as unknown type of noises, and at the same time also works comparably well against the model whose data fitting term is designed for a specific given noise type. While most of the existing methods for image restorations are designed specifically for a given type of noise, our model appears to be the first universal one for handling image restoration with various mixed noises and unknown type of noises. Extensive simulations on synthetic data show that our method is effective and robust in restoring images contaminated by additive Gaussian noise, Poisson noise, random-valued impulse noise, multiplicative Gamma noise and mixtures of these noises. Numerical results on real data show that it can remove noises without any prior knowledge of the noise distribution.