download0 view190
twitter facebook

공공누리This item is licensed Korea Open Government License

Title
Image denoising feedback framework using split Bregman approach
Author(s)
김정헌최광남Farhan Akram
Publication Year
2017-11-30
Abstract
In this paper, an image denoising feedback framework is proposed for both color and range images. The proposed method works on an error minimization principle using split Bregman method. At first image is denoised by computing means in the local neighborhood. The pixels that have big differences from the center of the local neighborhood compared to the noise variance are then extracted from the denoised image. There is a low correlation between the extracted pixels and their local neighborhood. This information is fed to the feedback function and denoising is performed again, iteratively, to minimize the error. In most cases, the proposed framework yields best results both qualitatively and quantitatively. It shows better denoising results than the bilateral filtering when the edge information in the input images is affected by intense noise. Moreover, during the denoising process feedback function ensures that the edges are not over smoothed. The proposed framework is applied to denoise both color and range images, which shows it works effectively on a wide variety of images unlike the evaluated state-of-the-art denoising methods.
Keyword
Image denoising; Split Bregman; Bilateral filtering; Image derivative; Color image; Range image
Journal Title
Expert Systems with Applications
Citation Volume
87
ISSN
0957-4174
Files in This Item:
There are no files associated with this item.
Appears in Collections:
7. KISTI 연구성과 > 학술지 발표논문
URI
https://repository.kisti.re.kr/handle/10580/14618
http://www.ndsl.kr/ndsl/search/detail/article/articleSearchResultDetail.do?cn=NART78162296
Export
RIS (EndNote)
XLS (Excel)
XML

Browse