An Alternative Variational Framework for Image Denoising

Figure 2

Lena image (300×300). (a) Original image. (b) Noisy image corrupted by Gaussian noise for 𝜎=30. (c) PM algorithm, 𝐾=5, 𝜏=0.25 (113 steps). (d) TV method, 𝜏=0.25 (129 steps). (e) Our method, 𝐾=0.1,  𝜏=5 (15 steps). (f) D-𝛼-PM algorithm, 𝐾=4,  𝜏=2 (10 steps).

(a)

(b)

(c)

(d)

(e)

(f)