&nbsp;基于空间域与频域特征自适应融合和类间边界区域增强的三维海马分割

白贺; 滕野; 冯蕾; 孟海伟; 汤煜春; 刘树伟

doi:10.16098/j.issn.0529-1356.2024.01.011

PDF(1877 KB)

解剖学报 ›› 2024, Vol. 55 ›› Issue (1) : 73-81. DOI: 10.16098/j.issn.0529-1356.2024.01.011

解剖学

基于空间域与频域特征自适应融合和类间边界区域增强的三维海马分割

白贺^1,2 滕野^1,2 冯蕾^1,2孟海伟^1,2汤煜春^1,2刘树伟^1,2*

作者信息 +

3D hippocampal segmentation based on spatial and frequency domain features adaptive fusion and inter-class boundary region enhancement

BAI He^1,2 TENG Ye^1,2 FENG Lei^1,2 MENG Hai-wei^1,2 TANG Yu-chun^1,2 LIU Shu-wei^1,2*

Author information +

文章历史 +

摘要

目的海马萎缩是诊断阿尔茨海默病等诸多精神疾病的临床重要标志，因此准确分割海马是一个重要的科学问题。随着深度学习的发展，人们提出了大量先进的自动分割方法。然而，由于MRI中各种噪声的影响以及海马不同类别之间不清晰的边界，三维海马分割仍然具有挑战性。因此本文旨在提出新的自动分割方法来更精确地分割海马头、体、尾。方法为了克服这些挑战，本文提出了两个策略。一种是空间域与频域特征自适应融合策略，通过快速傅立叶变换和卷积自动选择合适的频率组合，减少噪声对特征提取的影响。另一种是类间边界区域增强策略，它允许网络通过加权每个类之间边界区域的损失函数来增强对边界区域的学习，以达到精确定位边界和调节海马头、体、尾大小的目的。结果在50例青少年大脑MRI数据集上进行的实验表明，我们的方法实现了较先进的海马分割，海马头、体、尾相较于现有的方法都取得了一定的提升。消融实验证明我们提出的两种策略有效，我们还在260例Task04_Hippocampus数据集上验证了网络具有强大的泛化能力，说明本文提出的方法可用于更多的海马分割场景。结论我们提出的方法可以帮助临床医生更清楚地观测海马萎缩，并完成更精确的病情诊断和追踪。

Abstract

Objective Hippocampal atrophy is a clinically important marker for the diagnosis of many psychiatric disorders such as Alzheimer’s disease, so accurate segmentation of the hippocampus is an important scientific issue. With the development of deep learning, a large number of advanced automatic segmentation method have been proposed. However, 3D hippocampal segmentation is still challenging due to the effects of various noises in MRI and unclear boundaries between various classes of the hippocampus. Therefore, the aim of this paper is to propose new method to segment the hippocampal head, body, and tail more accurately. Methods To overcome these challenges, this paper proposed two strategies. One was the spatial and frequency domain features adaptive fusion strategy, which reduced the influence of noise on feature extraction by automatically selecting the appropriate frequency combination through fast Fourier transform and convolution. The other was an inter-class boundary region enhancement strategy, which allowed the network to focus on learning the boundary regions by weighting the loss function of the boundary regions between each class to achieve the goal of pinpointing the boundaries and regulating the size of the hippocampal head, body and tail.

Results Experiments performed on a 50-case teenager brain MRI dataset show that our method achieves state-of-the-art hippocampal segmentation. Hippocampal head, body and tail had been improved compared to the existing method. Ablation experiments demonstrated the effectiveness of our two proposed strategies, and we also validated that the network had a strong generalization ability on a 260-case Task04_Hippocampus dataset. It was shown that the method proposed in this paper could be used in more hippocampal segmentation scenarios. Conclusion The method proposed in this paper can help clinicians to observe hippocampal atrophy more clearly and accomplish more accurate diagnosis and follow-up of the condition.

导出引用

白贺滕野冯蕾孟海伟汤煜春刘树伟. 基于空间域与频域特征自适应融合和类间边界区域增强的三维海马分割[J]. 解剖学报. 2024, 55(1): 73-81 https://doi.org/10.16098/j.issn.0529-1356.2024.01.011

BAI He TENG Ye FENG Lei MENG Hai-wei1 TANG Yu-chun LIU Shu-wei. 3D hippocampal segmentation based on spatial and frequency domain features adaptive fusion and inter-class boundary region enhancement[J]. Acta Anatomica Sinica. 2024, 55(1): 73-81 https://doi.org/10.16098/j.issn.0529-1356.2024.01.011

中图分类号： R445.2

参考文献

［1］Tatulian SA. Challenges and hopes for Alzheimer’s disease［J］. Drug Discov Today, 2022, 27(4): 1027-1043.

［2］Schr?der J, Pantel J. Neuroimaging of hippocampal atrophy in early recognition of Alzheimer′ s disease-a critical appraisal after two decades of research［J］. Psychiatry Res Neuroimaging, 2016, 247: 71-78.

［3］Cole J, Costafreda SG, McGuffin P, et al. Hippocampal atrophy in first episode depression: a meta-analysis of magnetic resonance imaging studies［J］. J Affect Disord, 2011, 134(1-3): 483-487.
［4］Frisoni GB, Jack Jr CR, Bocchetta M, et al. The EADC-ADNI harmonized protocol for manual hippocampal segmentation on magnetic resonance: Evidence of validity［J］. Alzheimers Dement, 2015, 11(2): 111-125.
［5］Boccardi M, Bocchetta M, Apostolova LG, et al. Delphi definition of the EADC-ADNI harmonized protocol for hippocampal segmentation on magnetic resonance［J］. Alzheimers Dement, 2015, 11(2): 126-138.
［6］Lisman J, Buzsáki G, Eichenbaum H, et al. Viewpoints: how the hippocampus contributes to memory, navigation and cognition［J］. Nat Neurosci, 2017, 20(11): 1434-1447.
［7］Apostolova LG, Zarow C, Biado K, et al. Relationship between hippocampal atrophy and neuropathology markers: a 7T MRI validation study of the EADC-ADNI harmonized hippocampal segmentation protocol［J］. Alzheimers Dement, 2015, 11(2): 139-150.
［8］Çi?ek Ö , Abdulkadir A, Lienkamp SS, et al. 3D U-Net: learning dense volumetric segmentation from sparse annotation［C］.Medical Image Computing and Computer-Assisted Intervention-MICCAI 2016: 19th International Conference, 2016: 424-432.
［9］Antonelli M, Reinke A, Bakas S, et al. The medical segmentation decathlon［J］. Nat Commun, 2022, 13(1): 4128.
［10］Myronenko A. 3D MRI brain tumor segmentation using autoencoder regularization［C］. Brainlesion 4th International Workshop, 2019: 311-320.
［11］Dolz J, Gopinath K, Yuan J, et al. HyperDense-Net: a hyper-densely connected CNN for multi-modal image segmentation［J］. IEEE Trans Med Imaging, 2018, 38(5): 1116-1126.
［12］Nguyen T, Hua BS, Le N. 3d-ucaps: 3d capsules unet for volumetric image segmentation［C］. Medical Image Computing and Computer Assisted Intervention-MICCAI 2021: 24th International Conference, 2021: 548-558.
［13］Jiang Y, Zhang Z, Qin S, et al. APAUNet: axis projection attention UNet for small target in 3D medical segmentation［C］. Proceedings of The Asian Conference on Computer Vision，2022: 283-298.
［14］Peiris H, Hayat M, Chen Z, et al. A robust volumetric transformer for accurate 3D tumor segmentation［C］. International Conference on Medical Image Computing and Computer-Assisted Intervention, 2022: 162-172.
［15］Hatamizadeh A, Tang Y, Nath V, et al. Unetr: transformers for 3D medical image segmentation［C］. Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, 2022: 574-584.
［16］Zhao X, Zhang P, Song F, et al. Prior attention network for multi-lesion segmentation in medical images［J］. IEEE Trans Med Imaging, 2022, 41(12): 3812-3823.
［17］Ho TW, Qi H, Lai F, et al. Brain tumor segmentation using U-net and edge contour enhancement［C］. Proceedings of the 2019 3rd International Conference on Digital Signal Processing，2019: 75-79.
［18］Zhang M, Yu F, Zhao J, et al. BEFD: Boundary enhancement and feature denoising for vessel segmentation［C］. Medical Image Computing and Computer Assisted Intervention-MICCAI 2020: 23rd International Conference, 2020: 775-785.
［19］Wang S, Jiang A, Li X, et al. DPBET: a dual-path lung nodules segmentation model based on boundary enhancement and hybrid transformer［J］. Comput Biol Med, 2022, 151（Pt B）: 106330.
［20］Nagarajan I, Lakshmi Priya GG. Removal of noise in MRI images using a block difference-based filtering approach［J］. Int J Imaging Syst Technol, 2020, 30(1): 203-215.
［21］ZiyadSR, Radha V, Vaiyapuri T. Noise removal in lung LDCT images by novel discrete wavelet-based denoising with adaptive thresholding technique［J］. IJEHMC, 2021, 12(5): 1-15.
［22］Huynh KM, Chang W T, Chung S H, et al. Noise mapping and removal in complex-valued multi-channel MRI via optimal shrinkage of singular values［C］. International Conference on Medical Image Computing and Computer-Assisted Intervention, 2021: 191-200.
［23］JI DF, MA ZhB. Double modality fusion between CT and MRI for human head based on surface anatomic characters［J］. Acta Anatomica Sinica,2019,50(5):638-644.(in Chinese)
季达峰,马忠宾.基于表面解剖特征的人头部计算机断层与磁共振图像双模态融合［J］.解剖学报,2019,50(5):638-644.