Hybrid transformer and convolution iteratively optimized pyramid network for brain large deformation image registration

Abstract In recent years, the pyramid-based encoder-decoder network architecture has become a popular solution to the problem of large deformation image registration due to its excellent multi-scale deformation field prediction ability. However, there are two main limitations in existing research: o...

Full description

Saved in:
Bibliographic Details
Main Authors: Xinxin Cui, Yuee Zhou, Caihong Wei, Guodong Suo, Fengqing Jin, Jianlan Yang
Format: Article
Language:English
Published: Nature Portfolio 2025-05-01
Series:Scientific Reports
Subjects:
Online Access:https://doi.org/10.1038/s41598-025-00403-w
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849312031736332288
author Xinxin Cui
Yuee Zhou
Caihong Wei
Guodong Suo
Fengqing Jin
Jianlan Yang
author_facet Xinxin Cui
Yuee Zhou
Caihong Wei
Guodong Suo
Fengqing Jin
Jianlan Yang
author_sort Xinxin Cui
collection DOAJ
description Abstract In recent years, the pyramid-based encoder-decoder network architecture has become a popular solution to the problem of large deformation image registration due to its excellent multi-scale deformation field prediction ability. However, there are two main limitations in existing research: one is that it over-focuses on the fusion of multi-layer deformation sub-fields on the decoding path, while ignoring the impact of feature encoders on network performance; the other is the lack of specialized design for the characteristics of feature maps at different scales. To this end, we propose an innovative hybrid Transformer and convolution iteratively optimized pyramid network for large deformation brain image registration. Specifically, four encoder variants are designed to study the impact of different structures on the performance of the pyramid registration network. Secondly, the Swin-Transformer module is combined with the convolution iterative strategy, and each layer of the decoder is carefully designed according to the semantic information characteristics of different decoding layers. Extensive experimental results on three public brain magnetic resonance imaging datasets show that our method has the highest registration accuracy compared with 9 cutting-edge registration methods, which fully verifies the effectiveness and application potential of our model design.
format Article
id doaj-art-b73cb466f9ee4b9c8fc42d57ce21c676
institution Kabale University
issn 2045-2322
language English
publishDate 2025-05-01
publisher Nature Portfolio
record_format Article
series Scientific Reports
spelling doaj-art-b73cb466f9ee4b9c8fc42d57ce21c6762025-08-20T03:53:12ZengNature PortfolioScientific Reports2045-23222025-05-0115111810.1038/s41598-025-00403-wHybrid transformer and convolution iteratively optimized pyramid network for brain large deformation image registrationXinxin Cui0Yuee Zhou1Caihong Wei2Guodong Suo3Fengqing Jin4Jianlan Yang5School of Medical Information Engineering, Gansu University of Traditional Chinese MedicineSchool of Medical Information Engineering, Gansu University of Traditional Chinese MedicineQuanzhou Orthopedic Traumatological Hospital of Fujian University of Traditional Chinese MedicineSchool of Medical Information Engineering, Gansu University of Traditional Chinese MedicineSchool of Medical Information Engineering, Gansu University of Traditional Chinese MedicineSchool of Medical Information Engineering, Gansu University of Traditional Chinese MedicineAbstract In recent years, the pyramid-based encoder-decoder network architecture has become a popular solution to the problem of large deformation image registration due to its excellent multi-scale deformation field prediction ability. However, there are two main limitations in existing research: one is that it over-focuses on the fusion of multi-layer deformation sub-fields on the decoding path, while ignoring the impact of feature encoders on network performance; the other is the lack of specialized design for the characteristics of feature maps at different scales. To this end, we propose an innovative hybrid Transformer and convolution iteratively optimized pyramid network for large deformation brain image registration. Specifically, four encoder variants are designed to study the impact of different structures on the performance of the pyramid registration network. Secondly, the Swin-Transformer module is combined with the convolution iterative strategy, and each layer of the decoder is carefully designed according to the semantic information characteristics of different decoding layers. Extensive experimental results on three public brain magnetic resonance imaging datasets show that our method has the highest registration accuracy compared with 9 cutting-edge registration methods, which fully verifies the effectiveness and application potential of our model design.https://doi.org/10.1038/s41598-025-00403-wEnhanced pyramid encoderConvolution iterative optimizationTransformerBrain MRI
spellingShingle Xinxin Cui
Yuee Zhou
Caihong Wei
Guodong Suo
Fengqing Jin
Jianlan Yang
Hybrid transformer and convolution iteratively optimized pyramid network for brain large deformation image registration
Scientific Reports
Enhanced pyramid encoder
Convolution iterative optimization
Transformer
Brain MRI
title Hybrid transformer and convolution iteratively optimized pyramid network for brain large deformation image registration
title_full Hybrid transformer and convolution iteratively optimized pyramid network for brain large deformation image registration
title_fullStr Hybrid transformer and convolution iteratively optimized pyramid network for brain large deformation image registration
title_full_unstemmed Hybrid transformer and convolution iteratively optimized pyramid network for brain large deformation image registration
title_short Hybrid transformer and convolution iteratively optimized pyramid network for brain large deformation image registration
title_sort hybrid transformer and convolution iteratively optimized pyramid network for brain large deformation image registration
topic Enhanced pyramid encoder
Convolution iterative optimization
Transformer
Brain MRI
url https://doi.org/10.1038/s41598-025-00403-w
work_keys_str_mv AT xinxincui hybridtransformerandconvolutioniterativelyoptimizedpyramidnetworkforbrainlargedeformationimageregistration
AT yueezhou hybridtransformerandconvolutioniterativelyoptimizedpyramidnetworkforbrainlargedeformationimageregistration
AT caihongwei hybridtransformerandconvolutioniterativelyoptimizedpyramidnetworkforbrainlargedeformationimageregistration
AT guodongsuo hybridtransformerandconvolutioniterativelyoptimizedpyramidnetworkforbrainlargedeformationimageregistration
AT fengqingjin hybridtransformerandconvolutioniterativelyoptimizedpyramidnetworkforbrainlargedeformationimageregistration
AT jianlanyang hybridtransformerandconvolutioniterativelyoptimizedpyramidnetworkforbrainlargedeformationimageregistration