Brain MR Image Segmentation Based on Spatial Self-attention Mechanism and Depth Feature Reconstruction

doi:10.12068/j.issn.1005-3026.2023.02.004

Abstract

Abstract: Accurate segmentation of brain tissue in MR images is a key step in clinical diagnosis， surgical planning and adjuvant treatment. Deep-learning shows great potential in various image segmentation tasks， and existing models do not have an effective way to summarize the relationship between long-distance pixels. In the network decoding stage， the features of different levels cannot be well integrated， resulting in the inability to accurately locate. To overcome the above problems， this paper proposes a brain MR image segmentation method based on spatial self-attention mechanism and depth feature reconstruction， and constructs a 2D model that can fuse 3D information， which can quickly and accurately perform dense prediction on 3D structural images. The proposed method is fully experimented on MRBrainS13 data sets and IBSR data sets， and the results show that the model outperforms the current 2D model in 3D multimodal and unimodal brain MR image segmentation， with less computing and inference time compared to the 3D model， whereas the performance is very close.

Key words: brain image segmentation; fully convolutional network; spatial self-attention; channel attention; depth feature reconstruction

CLC Number:

TP20

WEI Ying， LIN Zi-han， QI Lin， LI Bo-qun. Brain MR Image Segmentation Based on Spatial Self-attention Mechanism and Depth Feature Reconstruction[J]. Journal of Northeastern University(Natural Science), 2023, 44(2): 177-185.

References

[1]Nuzziello N，Ciaccia L，Liguori M.Precision medicine in neurodegenerative diseases: some promising tips coming from the microRNAs′ world［J］.Cells，2020，9(1).DOI:10.3390/cells9010075.
[2]Akkus Z，Galimzianova A，Hoogi A，et al.Deep learning for brain MRI segmentation: state of the art and future directions［J］.Journal of Digital Imaging，2017，30(4):449-459.
[3]Minaee S，Boykov Y Y，Porikli F，et al.Image segmentation using deep learning: a survey［J］.IEEE Transactions on Pattern Analysis and Machine Intelligence，2022，44(7):523-3542.
[4]Devunooru S，Alsadoon A，Chandana P W C，et al.Deep learning neural networks for medical image segmentation of brain tumours for diagnosis: a recent review and taxonomy［J］.Journal of Ambient Intelligence and Humanized Computing，2021，12(1):455-483.
[5]Moeskops P，Viergever M A，Mendrik A M，et al.Automatic segmentation of MR brain images with a convolutional neural network［J］.IEEE Transactions on Medical Imaging，2016，35(5):1252-1261.
[6]Chen H，Dou Q，Yu L，et al.VoxResNet: deep voxelwise residual networks for brain segmentation from 3D MR images［J］.NeuroImage，2018，170: 446-455.
[7]Sun K，Xiao B，Liu D，et al.Deep high-resolution representation learning for human pose estimation［C］//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition.Long Beach，2019:5693-5703.
[8]Gu R，Wang G，Song T，et al.CA-Net: comprehensive attention convolutional neural networks for explainable medical image segmentation［J］.IEEE Transactions on Medical Imaging，2020，40(2):699-711.
[9]Lyksborg M，Puonti O，Agn M，et al.An ensemble of 2D convolutional neural networks for tumor segmentation［C］// Scandinavian Conference on Image Analysis.New York: Springer，2015:201-211.
[10]Long J，Shelhamer E，Darrell T.Fully convolutional networks for semantic segmentation［J］.IEEE Transactions on Pattern Analysis & Machine Intelligence，2017，39(4): 640-651.
[11]Ronneberger O，Fischer P，Brox T.U-Net: convolutional networks for biomedical image segmentation［C］//International Conference on Medical Image Computing & Computer-assisted Intervention.Munich，2015.DOI:10.1007/978-3-319-24574-4_28.
[12]Dolz J，Desrosiers C，Ayed I B.3D fully convolutional networks for subcortical segmentation in MRI: a large-scale study［J］.NeuroImage，2017，170:456-470.
[13]Cao X，Lin Y.Caggnet: crossing aggregation network for medical image segmentation［C］//2020 25th International Conference on Pattern Recognition(ICPR).Milan:IEEE，2020:1744-1750.
[14]Isensee F，Jaeger P F，Kohl S A A，et al.U-Net: a self-configuring method for deep learning-based biomedical image segmentation［J］.Nature Methods，2021，18(2):203-211.
[15]Chen L C，Papandreou G，Kokkinos I，et al.DeepLab: semantic image segmentation with deep convolutional nets，atrous convolution，and fully connected CRFs［J］.IEEE Transactions on Pattern Analysis and Machine Intelligence，2018，40(4):834-848.
[16]Francesco V，Adriana R，Kyunghyun C，et al. ReSeg: a recurrent neural network-based model for semantic segmentation［C］// Proceedings of 29th IEEE Conference on Computer Vision and Pattern Recognition Workshops.Las Vegas，2016: 426-433.
[17]Sérgio P，Pinto A，Oliveira J，et al.Automatic brain tissue segmentation in MR images using random forests and conditional random fields［J］.Journal of Neuroscience Methods，2016，270:111-123.
[18]Rajchl M，Baxter J S，Mcleod A J，et al.Hierarchical max-flow segmentation framework for multi-atlas segmentation with Kohonen self-organizing map based Gaussian mixture modeling［J］.Medical Image Analysis，2016:27:45-56.
[19]Andermatt S，Pezold S，Cattin P.Multi-dimensional gated recurrent units for the segmentation of biomedical 3D-data［M］.Cham: Springer International Publishing，2016.https://doi.org/10.1007/978-3-319-46976-8_15.
[20]Stollenga M F，Byeon W，Liwicki M，et al.Parallel multi-dimensional LSTM with application to fast biomedical volumetric image segmentation［C］//29th Annual Conference on Neural Information Processing Systems(NIPS).Montreal，2015:464-450.
[21]Valverde S，Oliver A，Cabezas M，et al.Comparison of 10 brain tissue segmentation methods using revisited IBSR annotations［J］.Journal of Magnetic Resonance Imaging，2015，41(1):93-101.