1 Introduction
2 Related works
2.1 Image prior and edge prediction
2.2 Convolutional neural networks
2.3 Multi-task learning
3 Methods
3.1 Multi-task learning
3.2 Multi-scale face deblurring network
3.3 Synthesis loss function
3.3.1 Content loss
3.3.2 Perceptual loss
3.3.3 Adversarial loss
3.3.4 Overall loss
4 Experimental
4.1 Implementation detail
4.2 Datasets
4.3 Ablation study
4.3.1 Multi-scale learning
Model | Loss | Helen | CelebA | ||
---|---|---|---|---|---|
PSNR | SSIM | PSNR | SSIM | ||
Multi-scale (1 scale) | Content loss | 22.70 | 0824 | 22.19 | 0.844 |
Multi-scale (2 scales) | Content loss | 23.30 | 0.847 | 22.57 | 0.856 |
Multi-task (2 scales) | Content loss |
23.69
|
0.852
|
23.04
|
0.859
|
4.3.2 Transfer learning
4.3.3 Additional synthesis analysis
Approach | Helen | CelebA | ||
---|---|---|---|---|
PSNR | SSIM | PSNR | SSIM | |
Content loss | 23.69 | 0.852 | 23.04 | 0.859 |
+ Perceptual loss | 24.21 | 0.857 | 23.48 | 0.864 |
+ Adversarial loss |
24.25
|
0.864
|
23.56
|
0.871
|
5 Result and discussion
5.1 Comparisons with state-of-the-arts
5.2 Face recognition
Method | Detection (%) | Top 1 (%) | Top 3 (%) | Top 5 (%) |
---|---|---|---|---|
Clear images | 100 | 71 | 84 | 89 |
Blurred images | 77.4 | 29.1 | 43.4 | 51.3 |
Krishnan et al. [17] | 80.0 | 33.8 | 48.9 | 56.6 |
Pan et al. [37] | 78.9 | 42.0 | 55.7 | 62.2 |
Shan et al. [2] | 76.0 | 32.4 | 46.9 | 54.0 |
Xu et al. [18] | 82.5 | 41.1 | 55.4 | 62.1 |
Cho and Lee [6] | 52.2 | 17.2 | 27.3 | 32.5 |
Zhong et al. [38] | 69.5 | 27.6 | 41.6 | 48.5 |
Nah et al. [15] | 86.0 | 40.1 | 55.3 | 62.4 |
Ours |
92
|
55
|
69
|
75
|
5.3 Execution time
Method | Implementation | CPU / GPU | S |
---|---|---|---|
Krishnan et al. [17] | MATLAB | CPU | 2.52 |
Pan et al. [37] | MATLAB | CPU | 8.11 |
Shan et al. [2] | C++ | CPU | 16.32 |
Xu et al. [18] | C++ | CPU | 0.31 |
Cho and Lee [6] | C++ | CPU | 0.41 |
Zhong et al. [38] | MATLAB | CPU | 8.07 |
Nah et al. [15] | MATLAB | GPU | 0.09 |
Ours | PYTHON | GPU |
0.02
|