The Research Of Image Denoising Based On Dictionary

With the advent of information age, digital products increase quickly and largely, high image quality are demanded. However, in the process of the images’ acquisition, transmission, reception, it is inevitable to subject to external and internal environment so that product noise, result in the decreased image quality. Not only the image’s reso-lution decreased, at the same time, but also sophisticated structure of image itself will be destroyed. Actually, these precise structure often play a Important role in applica-tion. That is bad for the following image processing. Therefore, suppressing noise effectively is crucial in the image processing. In this paper, Two kinds of image noise reduction methods are proposed.In the past decade or so, the concept of redundant dictionary and sparse repre-sentation had been used widely in image denoising, which results in excellent results. However, there is some disadvantages:1) many existing denoising method is limited in dictionary selection. The selection of the dictionary is usually dependent on human intervention. Usually, it is necessary to set a threshold artificially, which increase the uncertainty and complexity of the algorithm. Taking this point into account, the pa-per propose a denoising algorithm based on adaptive-dictionary. The complexity of the method is reduced effectively.2) In denoising yeild, as we know, how to balance the re-lationship between removing noise and containing the image’s details is very important. Denoising is a process of removing information. Many denoising algorithms remove noise, while a lot of meaningful details of the image itself can’t be well maintained. To preserve details of noised image, therefore, local structure information of image which is described by the steering kernel is considered in this paper. Various experiments have been accomplished and prove the two methods proposed in this paper to be effective in balancing the noise removing and the image details preserving.