ADMM algorithm for demosaicking deblurring denoising

Daniele Graziani (1) , Laure Blanc-Féraud (1) , Gilles Aubert (2)

1 MORPHEME - Morphologie et Images
2 LJAD - Laboratoire Jean Alexandre Dieudonné

The paper is concerned with the problem of demosaicking, deblurring and denoising a color image in the same time. The global model of the acquisition chain for a color image contains these three effects, then doing restoration in the same time as demosaicking make sense. We propose to take into account for correlation of spectral bands (R,G,B colors) by minimizing a criterion written in a nearly decorrelated basis. Then we adapt the Alternating Direction Multipliers Minimozation (ADMM) method to get the solution.

Mots clés: demosaicking; deburring; denoising ADMM algorithm

Domaines: Analyse numérique [math.NA]; Traitement du signal et de l'image [eess.SP]; Traitement du signal et de l'image [eess.SP]

Daniele Graziani, Laure Blanc-Féraud, Gilles Aubert. ADMM algorithm for demosaicking deblurring denoising. 2012. ⟨hal-00874610⟩