Color Image Denoising/Decomposition with Coupled Vectorial Total Variation


Abstract:

We study a regularization algorithm for color/vectorial images based on the vectorial total variation approach along with channel coupling for color image processing which facilitates the modeling of inter channel relations in multidimensional image data. We focus on penalizing channel gradient magnitude similarities by using L2 differences, which allow us to couple all the channels along with a vectorial total variation regularization for edge preserving smoothing of multi-channel images. A detailed mathematical analysis of the coupled vectorial total variation is provided. We are interested of applying our model to color image processing and in particular to denoising and decomposition. A fast global minimization based on the dual formulation of the total variation is used in our implementations which provides good decomposition and denoising results. Comparison with previous color image decomposition and denoising models are provided to demonstrate the advantages of our scheme.

Related projects: CMAC-Coupled Multiphase Contours, MultiAD

Example Decomposition Result

Top row: Cartoon images, Bottom row: Texture images

Reference:

J. C. Moreno, V. B. S. Prasath, J. Neves. Color image processing by vectorial total variation with gradient channels coupling. Inverse Problems and Imaging, vol 10, no 2, 461-497, May 2016. doi:10.3934/ipi.2016008

For anisotropic diffusion of mono channel (gray-scale images), please visit this project page.