TU Berlin

Communication Systems GroupScientific Publications

Page Content

to Navigation

Scientific Publications

Color Image Noise Reduction using Perceptual Maximum Variation Modeling for Color Diffusion
Citation key 0976Kim2006
Author Jang-Heon Kim and Thomas Sikora
Title of Book 7th International Workshop on Image Analysis for Multimedia Interactive Services (WIAMIS 2006)
Year 2006
Address Incheon, South Korea
Month apr
Abstract Diffusion is an efficient localized image regularization method based on the analysis of image structures such as direction and magnitude. However, the localization at weak features which have small brightness variations is fundamentally difficult. This often results in removal of weak features. We address this problem with perceptual maximum variation modeling. In our method, diffusion flow of color images is performed by evaluating the perceptual maximum variations which combine the small differences in both brightness and chromaticity, using a least squares optimization with principal component analysis (PCA). A consistency constraint is employed to avoid influence from global color distributions and to enhance homogeneous color regions. We apply our approach for de-noising of color images and obtain excellent improvements over existing methods.
Link to publication Download Bibtex entry


Quick Access

Schnellnavigation zur Seite über Nummerneingabe