The Successive Next Network as Augmented Regularization for Deformable Brain MR Image Registration
Deep-learning-based registration methods can not only save time but also automatically extract deep features from images. In order to obtain better registration performance, many scholars use cascade networks to realize a coarse-to-fine registration progress. However, such cascade networks will incr...
| Main Authors: | , , , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-03-01
|
| Series: | Sensors |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1424-8220/23/6/3208 |
| _version_ | 1797609062185041920 |
|---|---|
| author | Meng Li Shunbo Hu Guoqiang Li Fuchun Zhang Jitao Li Yue Yang Lintao Zhang Mingtao Liu Yan Xu Deqian Fu Wenyin Zhang Xing Wang |
| author_facet | Meng Li Shunbo Hu Guoqiang Li Fuchun Zhang Jitao Li Yue Yang Lintao Zhang Mingtao Liu Yan Xu Deqian Fu Wenyin Zhang Xing Wang |
| author_sort | Meng Li |
| collection | DOAJ |
| description | Deep-learning-based registration methods can not only save time but also automatically extract deep features from images. In order to obtain better registration performance, many scholars use cascade networks to realize a coarse-to-fine registration progress. However, such cascade networks will increase network parameters by an n-times multiplication factor and entail long training and testing stages. In this paper, we only use a cascade network in the training stage. Unlike others, the role of the second network is to improve the registration performance of the first network and function as an augmented regularization term in the whole process. In the training stage, the mean squared error loss function between the dense deformation field (DDF) with which the second network has been trained and the zero field is added to constrain the learned DDF such that it tends to 0 at each position and to compel the first network to conceive of a better deformation field and improve the network’s registration performance. In the testing stage, only the first network is used to estimate a better DDF; the second network is not used again. The advantages of this kind of design are reflected in two aspects: (1) it retains the good registration performance of the cascade network; (2) it retains the time efficiency of the single network in the testing stage. The experimental results show that the proposed method effectively improves the network’s registration performance compared to other state-of-the-art methods. |
| first_indexed | 2024-03-11T05:55:20Z |
| format | Article |
| id | doaj.art-6e2016d31ab142cb8481c34f46b268cb |
| institution | Directory Open Access Journal |
| issn | 1424-8220 |
| language | English |
| last_indexed | 2024-03-11T05:55:20Z |
| publishDate | 2023-03-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Sensors |
| spelling | doaj.art-6e2016d31ab142cb8481c34f46b268cb2023-11-17T13:47:28ZengMDPI AGSensors1424-82202023-03-01236320810.3390/s23063208The Successive Next Network as Augmented Regularization for Deformable Brain MR Image RegistrationMeng Li0Shunbo Hu1Guoqiang Li2Fuchun Zhang3Jitao Li4Yue Yang5Lintao Zhang6Mingtao Liu7Yan Xu8Deqian Fu9Wenyin Zhang10Xing Wang11School of Information Science and Engineering, Linyi University, Linyi 276000, ChinaSchool of Information Science and Engineering, Linyi University, Linyi 276000, ChinaSchool of Information Science and Engineering, Linyi University, Linyi 276000, ChinaSchool of Information Science and Engineering, Linyi University, Linyi 276000, ChinaSchool of Information Science and Engineering, Linyi University, Linyi 276000, ChinaSchool of Information Science and Engineering, Linyi University, Linyi 276000, ChinaSchool of Information Science and Engineering, Linyi University, Linyi 276000, ChinaSchool of Information Science and Engineering, Linyi University, Linyi 276000, ChinaSchool of Information Science and Engineering, Linyi University, Linyi 276000, ChinaSchool of Information Science and Engineering, Linyi University, Linyi 276000, ChinaSchool of Information Science and Engineering, Linyi University, Linyi 276000, ChinaSchool of Information Science and Engineering, Linyi University, Linyi 276000, ChinaDeep-learning-based registration methods can not only save time but also automatically extract deep features from images. In order to obtain better registration performance, many scholars use cascade networks to realize a coarse-to-fine registration progress. However, such cascade networks will increase network parameters by an n-times multiplication factor and entail long training and testing stages. In this paper, we only use a cascade network in the training stage. Unlike others, the role of the second network is to improve the registration performance of the first network and function as an augmented regularization term in the whole process. In the training stage, the mean squared error loss function between the dense deformation field (DDF) with which the second network has been trained and the zero field is added to constrain the learned DDF such that it tends to 0 at each position and to compel the first network to conceive of a better deformation field and improve the network’s registration performance. In the testing stage, only the first network is used to estimate a better DDF; the second network is not used again. The advantages of this kind of design are reflected in two aspects: (1) it retains the good registration performance of the cascade network; (2) it retains the time efficiency of the single network in the testing stage. The experimental results show that the proposed method effectively improves the network’s registration performance compared to other state-of-the-art methods.https://www.mdpi.com/1424-8220/23/6/3208brain image registrationgeneration adversarial networkdeep learning |
| spellingShingle | Meng Li Shunbo Hu Guoqiang Li Fuchun Zhang Jitao Li Yue Yang Lintao Zhang Mingtao Liu Yan Xu Deqian Fu Wenyin Zhang Xing Wang The Successive Next Network as Augmented Regularization for Deformable Brain MR Image Registration Sensors brain image registration generation adversarial network deep learning |
| title | The Successive Next Network as Augmented Regularization for Deformable Brain MR Image Registration |
| title_full | The Successive Next Network as Augmented Regularization for Deformable Brain MR Image Registration |
| title_fullStr | The Successive Next Network as Augmented Regularization for Deformable Brain MR Image Registration |
| title_full_unstemmed | The Successive Next Network as Augmented Regularization for Deformable Brain MR Image Registration |
| title_short | The Successive Next Network as Augmented Regularization for Deformable Brain MR Image Registration |
| title_sort | successive next network as augmented regularization for deformable brain mr image registration |
| topic | brain image registration generation adversarial network deep learning |
| url | https://www.mdpi.com/1424-8220/23/6/3208 |
| work_keys_str_mv | AT mengli thesuccessivenextnetworkasaugmentedregularizationfordeformablebrainmrimageregistration AT shunbohu thesuccessivenextnetworkasaugmentedregularizationfordeformablebrainmrimageregistration AT guoqiangli thesuccessivenextnetworkasaugmentedregularizationfordeformablebrainmrimageregistration AT fuchunzhang thesuccessivenextnetworkasaugmentedregularizationfordeformablebrainmrimageregistration AT jitaoli thesuccessivenextnetworkasaugmentedregularizationfordeformablebrainmrimageregistration AT yueyang thesuccessivenextnetworkasaugmentedregularizationfordeformablebrainmrimageregistration AT lintaozhang thesuccessivenextnetworkasaugmentedregularizationfordeformablebrainmrimageregistration AT mingtaoliu thesuccessivenextnetworkasaugmentedregularizationfordeformablebrainmrimageregistration AT yanxu thesuccessivenextnetworkasaugmentedregularizationfordeformablebrainmrimageregistration AT deqianfu thesuccessivenextnetworkasaugmentedregularizationfordeformablebrainmrimageregistration AT wenyinzhang thesuccessivenextnetworkasaugmentedregularizationfordeformablebrainmrimageregistration AT xingwang thesuccessivenextnetworkasaugmentedregularizationfordeformablebrainmrimageregistration AT mengli successivenextnetworkasaugmentedregularizationfordeformablebrainmrimageregistration AT shunbohu successivenextnetworkasaugmentedregularizationfordeformablebrainmrimageregistration AT guoqiangli successivenextnetworkasaugmentedregularizationfordeformablebrainmrimageregistration AT fuchunzhang successivenextnetworkasaugmentedregularizationfordeformablebrainmrimageregistration AT jitaoli successivenextnetworkasaugmentedregularizationfordeformablebrainmrimageregistration AT yueyang successivenextnetworkasaugmentedregularizationfordeformablebrainmrimageregistration AT lintaozhang successivenextnetworkasaugmentedregularizationfordeformablebrainmrimageregistration AT mingtaoliu successivenextnetworkasaugmentedregularizationfordeformablebrainmrimageregistration AT yanxu successivenextnetworkasaugmentedregularizationfordeformablebrainmrimageregistration AT deqianfu successivenextnetworkasaugmentedregularizationfordeformablebrainmrimageregistration AT wenyinzhang successivenextnetworkasaugmentedregularizationfordeformablebrainmrimageregistration AT xingwang successivenextnetworkasaugmentedregularizationfordeformablebrainmrimageregistration |