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

Figure 4

Denoising and deconvolution with the Gamma TV via [117].

(a) Grayscale “Tank” image corrupted with Gamma noise (𝐿=10). SNR: −4.0 dB., SSIM: 0.189

(b) Blurred grayscale “Tank” image corrupted with Gamma noise (𝐿=10). SNR: −4.1 dB., SSIM: 0.101

(c) Blurred color “Lena” image corrupted with Gamma noise (𝐿=10). SNR: 2.3 dB

(d) Gamma-TV denoising. SNR: 8.87 dB, SSIM: 0.57

(e) Gamma-TV deconvolution. SNR: 8.05 dB, SSIM: 0.50

(f) Gamma-TV deconvolution. SNR: 14.66 dB