Fig. 4
From: Noise-resistant network: a deep-learning method for face recognition under noise
The sample images of AR database injected with different noise. a–d The first line of Fig. 4 indicates the samples injected with Gaussian noise σ = 0.05, 0.10, 0.15, 0.20. e–h The second line indicates the samples injected with uniform noise p = 0.10, 0.20, 0.40, 0.70. i–l The third line indicates the samples injected with salt and pepper noise d = 0.05, 0.10, 0.15, 0.25.