Laser & Optoelectronics Progress, Volume. 57, Issue 16, 161009(2020)
Nonrigid Image Registration Algorithm Based on Mutual Information Active Demons
Figures & Tables(11)
Fig. 3. brain registration results of different algorithms. (a) Original image; (b) image to be registered; (c) Demons; (d) Active Demons(α=2); (e) proposed algorithm; (f) TV-L1; (g) Spectral Log-Demons; (h) CDMFFD
Fig. 4. Residual images between brain registration results and original image. (a) Initial difference; (b) Demons; (c) Active Demons(α=2); (d) proposed algorithm; (e) TV-L1; (f) Spectral Log-Demons; (g) CDMFFD
Fig. 5. Lena registration results of different algorithms. (a) Original image; (b) image to be registered; (c) Demons; (d) Active Demons (α=1.4); (e) proposed algorithm; (f) TV-L1; (g) Spectral Log-Demons; (h) CDMFFD
Fig. 6. Residual images between Lena registration results and original image. (a) Initial difference; (b) Demons; (c) Active Demons (α=1.4); (d) proposed algorithm; (e) TV-L1; (f) Spectral Log-Demons; (g) CDMFFD
Fig. 7. checkboard registration results of different algorithms. (a) Original image; (b) image to be registered; (c) Demons; (d) Active Demons (α=0.2); (e) proposed algorithm; (f) TV-L1; (g) Spectral Log-Demons; (h) CDMFFD
Fig. 8. Residual images between checkboard registration results and original image. (a) Initial difference; (b) Demons; (c) Active Demons (α=0.2); (d) proposed algorithm; (e) TV-L1; (f) Spectral Log-Demons; (g) CDMFFD
Table 1. Objective analysis and comparison of brain registration results of different algorithms
View tableTable 1. Objective analysis and comparison of brain registration results of different algorithms
Algorithm PSNR /dB SSIM CC RM/10-4 RV Before registration 24.5711 0.7043 0.8594 3.0280 89.0000 Demons 35.6029 0.9607 0.9889 0.8367 11.0000 Active Demons 41.2332 0.9844 0.9970 1.2711 7.5000 Proposed 44.5809 0.9910 0.9986 0.0095 0.0034 TV-L1 50.6751 0.9923 0.9996 0.1593 0.0850 Spectral Log-Demons 25.0769 0.9232 0.9973 0.0434 0.0012 CDMFFD 25.2939 0.9459 0.9988 0.0405 0.0013
Table 2. Objective analysis and comparison of Lena registration results of different algorithms
View tableTable 2. Objective analysis and comparison of Lena registration results of different algorithms
Algorithm PSNR /dB SSIM CC RM/10-4 RV Before registration 20.4835 0.6385 0.7487 3.0280 89.0000 Demons 28.9179 0.8942 0.9635 8.3670 11.0000 Active Demons 30.9851 0.9176 0.9774 0.4700 7.3000 Proposed 32.6490 0.9318 0.9847 0.2335 0.0005 TV-L1 43.4070 0.9914 0.9987 0.5989 0.4500 Spectral Log-Demons 19.8777 0.7974 0.9622 0.0270 0.0096 CDMFFD 19.4924 0.8625 0.9880 0.0241 0.0107
Table 3. Objective analysis and comparison of checkboard registration results of different algorithms
View tableTable 3. Objective analysis and comparison of checkboard registration results of different algorithms
Algorithm PSNR /dB SSIM CC RM/10-4 RV Before registration 8.0089 0.3969 0.6749 1.1781 0.1582 Demons 22.7378 0.9480 0.9898 3.0170 0.0052 Active Demons 23.1830 0.9525 0.9908 12.0000 0.0048 Proposed 24.2339 0.9533 0.9910 7.5285 0.0047 TV-L1 32.9810 0.9993 0.9989 0.0602 0.0005 Spectral Log-Demons 12.6212 0.4220 0.8958 0.0015 0.0547 CDMFFD 26.9579 0.9783 0.9963 1.2134 0.0020
Tools
Get Citation
Copy Citation Text
Dan Zhang, Huan Huang, Zhenhong Shang. Nonrigid Image Registration Algorithm Based on Mutual Information Active Demons[J]. Laser & Optoelectronics Progress, 2020, 57(16): 161009
Paper Information
Category: Image Processing
Received: Dec. 4, 2019
Accepted: Jan. 6, 2020
Published Online: Aug. 5, 2020
The Author Email: Huan Huang (1427903561@qq.com)
© Copyright 2018-2021 | Chinese Laser Press.
All Rights Reserved 沪ICP备15018463号-20