Research‎ > ‎

MindX: Mixed Impulse Poisson-Gaussian Denoising


Description

This work aims at developing denoising algorithms for images corrupted by mixed impulse noise and Poisson-Gaussian noise. This usually arises from situations of low photon counts and limited illuminations. We develop a combinatorial algorithm based on proximal algorithms that seamlessly combines state of the art priors, such as TV, with other priors, e.g. BM3D. We compare to the state of the art methods and show superior performance.

References

Mohamed Aly and Wolfgang Heidrich. MindX: Denoising Mixed Impulse Poisson-Gaussian Noise Using Proximal Algorithms. arXiv preprint arXiv:1608.07802, 2016. [pdf]

Comments