Poisson noise removal is a very challenging problem since this kind of noise may contaminate images in a totally different way from additive noise. This research proposal is focused on the development of image restoration models for Poisson noise removal and deblurring. The importance of Poisson noise mainly comes from the fact that this kind noise can be commonly found in many real word image processing applications such as microscope images and medical images. Unlike additive noises, these noises are much more difficult to be removed from the corrupted image due to the fact that their statistical distributions are signal dependant. In this proposal, we devote to variational approaches based on the theory of image restoration, optimization and statistics. Indeed, We will first examine a fast method for a classical variational model revoked by Poisson noise and blurring. And then we derive a new variational method for Poisson noise removal and blurring. The main contribution here is that the variation principle and sparse representation for image processing are seamlessly combined together for possion noise removal and blurring. In addition, efficient method is also devised to effectively solve the proposed model. We then turn to extending the previous results to other types of noise such as impulse noise and multiplicative noise. Finally, we study the non-blind deconvultion problem under Poisson noise.
|Effective start/end date||1/01/12 → 31/12/14|
Explore the research topics touched on by this project. These labels are generated based on the underlying awards/grants. Together they form a unique fingerprint.