Journal of Electrical and Computer Engineering

Review Article

Total Variation Regularization Algorithms for Images Corrupted with Different Noise Models: A Review

Figure 2

Denoising with the -TV and the adaptive -TV methods (via [81] and via [41, 100], resp.) for the grayscale “Boats” ( ) image and the color “Barbara” ( ) image.
217021.fig.002a
(a) Grayscale “Boats” image corrupted with 50% Salt & Pepper noise
217021.fig.002b
(b) -TV denoising with regularization parameter (SNR: 9.01 dB, SSIM: 0.66)
217021.fig.002c
(c) Adaptive -TV denoising (SNR: 15.81 dB, SSIM: 0.89)
217021.fig.002d
(d) Color “Barbara” image corrupted with 80% Salt & Pepper noise
217021.fig.002e
(e) -TV denoising with regularization parameter (SNR: 4.52 dB)
217021.fig.002f
(f) Adaptive -TV denoising (SNR: 10.89 dB)