Enhancing unsupervised learning in medical image registration through scale-aware context aggregation
Summary: Deformable image registration (DIR) is essential for medical image analysis, facilitating the establishment of dense correspondences between images to analyze complex deformations. Traditional registration algorithms often require significant computational resources due to iterative optimiz...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2025-02-01
|
| Series: | iScience |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S2589004224029614 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832595843990945792 |
|---|---|
| author | Yuchen Liu Ling Wang Xiaolin Ning Yang Gao Defeng Wang |
| author_facet | Yuchen Liu Ling Wang Xiaolin Ning Yang Gao Defeng Wang |
| author_sort | Yuchen Liu |
| collection | DOAJ |
| description | Summary: Deformable image registration (DIR) is essential for medical image analysis, facilitating the establishment of dense correspondences between images to analyze complex deformations. Traditional registration algorithms often require significant computational resources due to iterative optimization, while deep learning approaches face challenges in managing diverse deformation complexities and task requirements. We introduce ScaMorph, an unsupervised learning model for DIR that employs scale-aware context aggregation, integrating multiscale mixed convolution with lightweight multiscale context fusion. This model effectively combines convolutional networks and vision transformers, addressing various registration tasks. We also present diffeomorphic variants of ScaMorph to maintain topological deformations. Extensive experiments on 3D medical images across five applications—atlas-to-patient and inter-patient brain magnetic resonance imaging (MRI) registration, inter-modal brain MRI registration, inter-patient liver computed tomography (CT) registration as well as inter-modal abdomen MRI-CT registration—demonstrate that our model significantly outperforms existing methods, highlighting its effectiveness and broader implications for enhancing medical image registration techniques. |
| format | Article |
| id | doaj-art-e9ba74c9f1a243f898ab82591e1449c2 |
| institution | Kabale University |
| issn | 2589-0042 |
| language | English |
| publishDate | 2025-02-01 |
| publisher | Elsevier |
| record_format | Article |
| series | iScience |
| spelling | doaj-art-e9ba74c9f1a243f898ab82591e1449c22025-01-18T05:05:05ZengElsevieriScience2589-00422025-02-01282111734Enhancing unsupervised learning in medical image registration through scale-aware context aggregationYuchen Liu0Ling Wang1Xiaolin Ning2Yang Gao3Defeng Wang4School of Instrumentation Science and Opto-electronics Engineering, Beihang University, Beijing 100191, ChinaInstitute of Large-Scale Scientific Facility and Centre for Zero Magnetic Field Science, Beihang University, Beijing 100191, ChinaSchool of Instrumentation Science and Opto-electronics Engineering, Beihang University, Beijing 100191, China; Institute of Large-Scale Scientific Facility and Centre for Zero Magnetic Field Science, Beihang University, Beijing 100191, China; Hefei National Laboratory, Hefei 230000, ChinaSchool of Instrumentation Science and Opto-electronics Engineering, Beihang University, Beijing 100191, China; Institute of Large-Scale Scientific Facility and Centre for Zero Magnetic Field Science, Beihang University, Beijing 100191, China; Hefei National Laboratory, Hefei 230000, China; Corresponding authorSchool of Instrumentation Science and Opto-electronics Engineering, Beihang University, Beijing 100191, China; Corresponding authorSummary: Deformable image registration (DIR) is essential for medical image analysis, facilitating the establishment of dense correspondences between images to analyze complex deformations. Traditional registration algorithms often require significant computational resources due to iterative optimization, while deep learning approaches face challenges in managing diverse deformation complexities and task requirements. We introduce ScaMorph, an unsupervised learning model for DIR that employs scale-aware context aggregation, integrating multiscale mixed convolution with lightweight multiscale context fusion. This model effectively combines convolutional networks and vision transformers, addressing various registration tasks. We also present diffeomorphic variants of ScaMorph to maintain topological deformations. Extensive experiments on 3D medical images across five applications—atlas-to-patient and inter-patient brain magnetic resonance imaging (MRI) registration, inter-modal brain MRI registration, inter-patient liver computed tomography (CT) registration as well as inter-modal abdomen MRI-CT registration—demonstrate that our model significantly outperforms existing methods, highlighting its effectiveness and broader implications for enhancing medical image registration techniques.http://www.sciencedirect.com/science/article/pii/S2589004224029614Medical imagingClinical neuroscienceBioinformatics |
| spellingShingle | Yuchen Liu Ling Wang Xiaolin Ning Yang Gao Defeng Wang Enhancing unsupervised learning in medical image registration through scale-aware context aggregation iScience Medical imaging Clinical neuroscience Bioinformatics |
| title | Enhancing unsupervised learning in medical image registration through scale-aware context aggregation |
| title_full | Enhancing unsupervised learning in medical image registration through scale-aware context aggregation |
| title_fullStr | Enhancing unsupervised learning in medical image registration through scale-aware context aggregation |
| title_full_unstemmed | Enhancing unsupervised learning in medical image registration through scale-aware context aggregation |
| title_short | Enhancing unsupervised learning in medical image registration through scale-aware context aggregation |
| title_sort | enhancing unsupervised learning in medical image registration through scale aware context aggregation |
| topic | Medical imaging Clinical neuroscience Bioinformatics |
| url | http://www.sciencedirect.com/science/article/pii/S2589004224029614 |
| work_keys_str_mv | AT yuchenliu enhancingunsupervisedlearninginmedicalimageregistrationthroughscaleawarecontextaggregation AT lingwang enhancingunsupervisedlearninginmedicalimageregistrationthroughscaleawarecontextaggregation AT xiaolinning enhancingunsupervisedlearninginmedicalimageregistrationthroughscaleawarecontextaggregation AT yanggao enhancingunsupervisedlearninginmedicalimageregistrationthroughscaleawarecontextaggregation AT defengwang enhancingunsupervisedlearninginmedicalimageregistrationthroughscaleawarecontextaggregation |