Figure 10From: An augmented Lagrangian multi-scale dictionary learning algorithmThe denoised result of image "Cameraman" with σ = 15. Top plots: (a) the initial dictionary, (b) the dictionary trained by K-SVD, and (c) the dictionary trained by ALM-DL algorithm. Bottom plots (d) the reference image, (e) denoised image by K-SVD and (f) our ALM-DL algorithm.Back to article page