《计算机应用》唯一官方网站 ›› 2024, Vol. 44 ›› Issue (9): 2911-2918.DOI: 10.11772/j.issn.1001-9081.2023091332
王熙源1, 张战成1(), 徐少康2, 张宝成3, 罗晓清4, 胡伏原1
收稿日期:
2023-09-27
修回日期:
2023-12-28
接受日期:
2024-01-08
发布日期:
2024-01-31
出版日期:
2024-09-10
通讯作者:
张战成
作者简介:
王熙源(1999—),男,江苏南京人,硕士研究生,CCF会员,主要研究方向:医学影像、图像配准基金资助:
Xiyuan WANG1, Zhancheng ZHANG1(), Shaokang XU2, Baocheng ZHANG3, Xiaoqing LUO4, Fuyuan HU1
Received:
2023-09-27
Revised:
2023-12-28
Accepted:
2024-01-08
Online:
2024-01-31
Published:
2024-09-10
Contact:
Zhancheng ZHANG
About author:
WANG Xiyuan, born in 1999, M. S. candidate. His research interests include medical image, image registration.Supported by:
摘要:
3D/2D配准是手术导航的关键技术,现有基于深度学习的配准方法通过网络提取图像特征,继而回归出相应的姿态变换参数。此类方法依赖于真实的样本以及对应的3D标签用于训练,然而这部分专家标注的医疗数据十分稀缺。替代的方案用数字重建放射影像(DRR)图像训练网络,由于图像特征跨域的差异,在X射线图像上难以保持原有的配准精度。针对上述问题,设计基于自注意力的无监督跨域迁移网络(UCDTN),无须依赖X射线图像与其对应的3D空间标签作为训练样本,将源域所捕获的图像特征与空间变换间的对应关系迁移到目标域,借助公共特征减小域间特征的差距、降低跨域所带来的负面影响。实验结果表明,UCDTN预测结果的平均配准误差(mTRE)为2.66 mm;与未经跨域迁移训练的模型相比,mTRE指标降低了70.61%,验证了UCDTN在跨域配准任务上的有效性。
中图分类号:
王熙源, 张战成, 徐少康, 张宝成, 罗晓清, 胡伏原. 面向手术导航3D/2D配准的无监督跨域迁移网络[J]. 计算机应用, 2024, 44(9): 2911-2918.
Xiyuan WANG, Zhancheng ZHANG, Shaokang XU, Baocheng ZHANG, Xiaoqing LUO, Fuyuan HU. Unsupervised cross-domain transfer network for 3D/2D registration in surgical navigation[J]. Journal of Computer Applications, 2024, 44(9): 2911-2918.
序号 | 网络 | mTRE/mm | 训练样本 | |
---|---|---|---|---|
冠状面 | 矢状面 | |||
1 | ResNet(源域) | 5.42 | 4.75 | DRR |
2 | UCDTN(源域) | 1.32 | 1.06 | DRR |
3 | ResNet(跨域1) | 37.42 | 24.71 | DRR |
4 | UCDTN(跨域1) | 21.80 | 18.81 | DRR |
5 | ResNet(跨域2) | 14.33 | 12.57 | DRR,伪Xray |
6 | UCDTN(跨域2) | 9.71 | 8.39 | DRR,伪Xray |
7 | UCDTN(域迁移) | 3.18 | 2.13 | DRR,伪Xray |
表1 不同网络的有效性分析
Tab. 1 Effective analysis of different networks
序号 | 网络 | mTRE/mm | 训练样本 | |
---|---|---|---|---|
冠状面 | 矢状面 | |||
1 | ResNet(源域) | 5.42 | 4.75 | DRR |
2 | UCDTN(源域) | 1.32 | 1.06 | DRR |
3 | ResNet(跨域1) | 37.42 | 24.71 | DRR |
4 | UCDTN(跨域1) | 21.80 | 18.81 | DRR |
5 | ResNet(跨域2) | 14.33 | 12.57 | DRR,伪Xray |
6 | UCDTN(跨域2) | 9.71 | 8.39 | DRR,伪Xray |
7 | UCDTN(域迁移) | 3.18 | 2.13 | DRR,伪Xray |
感知深度d | ||
---|---|---|
1 | 8.21 | 0.023 |
2 | 6.08 | 0.016 |
3 | 4.62 | 0.007 |
4 | 5.33 | 0.009 |
表2 不同感知域深度的mTRE和MAE
Tab. 2 mTRE and MAE of for perceptual domains with different depths
感知深度d | ||
---|---|---|
1 | 8.21 | 0.023 |
2 | 6.08 | 0.016 |
3 | 4.62 | 0.007 |
4 | 5.33 | 0.009 |
实验序号 | 方法 | ||
---|---|---|---|
1 | 预训练( | 6.63 | 0.017 |
2 | 预训练( | 4.21 | 0.008 |
3 | 预训练( | 8.27 | 0.022 |
表3 辅助域的效果对比 (mm)
Tab. 3 Effectiveness comparison of auxiliary domain
实验序号 | 方法 | ||
---|---|---|---|
1 | 预训练( | 6.63 | 0.017 |
2 | 预训练( | 4.21 | 0.008 |
3 | 预训练( | 8.27 | 0.022 |
方法序号 | 损失函数 | ||
---|---|---|---|
1 | 19.84 | 0.117 | |
2 | 8.62 | 0.023 | |
3 | 4.43 | 0.018 | |
4 | 6.21 | 0.015 | |
5 | 2.74 | 0.005 |
表4 不同损失的表现 (mm)
Tab. 4 Performance of different losses
方法序号 | 损失函数 | ||
---|---|---|---|
1 | 19.84 | 0.117 | |
2 | 8.62 | 0.023 | |
3 | 4.43 | 0.018 | |
4 | 6.21 | 0.015 | |
5 | 2.74 | 0.005 |
方法 | 耗时/s | ||
---|---|---|---|
Opt-GO[ | 12.74 | 0.064 | 21.5 |
Opt-GC[ | 11.21 | 0.051 | 17.6 |
Opt-NGI[ | 12.29 | 0.062 | 23.1 |
姿态编码[ | 23.83 | 0.191 | 1.7 |
MLP[ | 32.17 | 0.244 | 3.1 |
姿态编码+Opt-GC | 8.73 | 0.027 | 8.6 |
UCDTN | 2.33 | 0.004 | 1.8 |
表5 不同方法的配准性能对比
Tab. 5 Registration performance comparison among different methods
方法 | 耗时/s | ||
---|---|---|---|
Opt-GO[ | 12.74 | 0.064 | 21.5 |
Opt-GC[ | 11.21 | 0.051 | 17.6 |
Opt-NGI[ | 12.29 | 0.062 | 23.1 |
姿态编码[ | 23.83 | 0.191 | 1.7 |
MLP[ | 32.17 | 0.244 | 3.1 |
姿态编码+Opt-GC | 8.73 | 0.027 | 8.6 |
UCDTN | 2.33 | 0.004 | 1.8 |
1 | MARKELJ P, TOMAŽEVIČ D, LIKAR B, et al. A review of 3D/2D registration methods for image-guided interventions [J]. Medical Image Analysis, 2012, 16(3): 642-661. |
2 | BRENNER D J, HALL E J. Computed tomography — an increasing source of radiation exposure [J]. The New England Journal of Medicine, 2007, 357(22): 2277-2284. |
3 | MACKIE T R, KAPATOES J, RUCHALA K, et al. Image guidance for precise conformal radiotherapy [J]. International Journal of Radiation Oncology, Biology, Physics, 2003, 56(1): 89-105. |
4 | UNBERATH M, GAO C, HU Y, et al. The impact of machine learning on 2D/3D registration for image-guided interventions: a systematic review and perspective [J]. Frontiers in Robotics and AI, 2021, 8: 716007. |
5 | VARNAVAS A, CARRELL T, PENNEY G. Increasing the automation of a 2D-3D registration system [J]. IEEE Transactions on Medical Imaging, 2013, 32(2): 387-399. |
6 | OUADAH S, JACOBSON M, STAYMAN J W, et al. Correction of patient motion in cone-beam CT using 3D-2D registration [J]. Physics in Medicine and Biology, 2017, 62(23): 8813-8831. |
7 | OTAKE Y, WANG A S, STAYMAN J W, et al. Robust 3D-2D image registration: application to spine interventions and vertebral labeling in the presence of anatomical deformation [J]. Physics in Medicine and Biology, 2013, 58(23): 8535-8553. |
8 | GOUVEIA A R, METZ C, FREIRE L, et al. Registration-by-regression of coronary CTA and X-ray angiography [J]. Computer Methods in Biomechanics and Biomedical Engineering: Imaging & Visualization, 2017, 5(3): 208-220. |
9 | C-R CHOU, FREDERICK B, MAGERAS G, et al. 2D/3D image registration using regression learning [J]. Computer Vision and Image Understanding, 2013, 117(9): 1095-1106. |
10 | MIAO S, WANG Z J, LIAO R. A CNN regression approach for real-time 2D/3D registration [J]. IEEE Transactions on Medical Imaging, 2016, 35(5): 1352-1363. |
11 | LIAO R, MIAO S, DE TOURNEMIRE P, et al. An artificial agent for robust image registration [J]. Proceedings of the AAAI Conference on Artificial Intelligence, 2017, 31(1): 4168-4175. |
12 | MIAO S, PIAT S, FISCHER P, et al. Dilated FCN for multi-agent 2D/3D medical image registration [J]. Proceedings of the AAAI Conference on Artificial Intelligence,2018, 32(1): 4694-4701. |
13 | 徐少康, 张战成, 姚浩男, 等. 基于姿态编码器的2D/3D脊椎医学图像实时配准方法[J]. 计算机应用, 2023, 43(2): 589-594. |
XU S K, ZHANG Z C, YAO H N, et al. 2D/3D spine medical image real-time registration method based on pose encoder[J]. Journal of Computer Applications, 2023, 43(2): 589-594. | |
14 | BALAKRISHNAN G, ZHAO A, SABUNCU M R, et al. VoxelMorph: a learning framework for deformable medical image registration [J]. IEEE Transactions on Medical Imaging, 2019, 38(8): 1788-1800. |
15 | MOK T C W, CHUNG A C S. Fast symmetric diffeomorphic image registration with convolutional neural networks [C]// Proceedings of the 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2020: 4643-4652. |
16 | 张家岗, 李达平, 杨晓东, 等. 基于深度卷积特征光流的形变医学图像配准算法 [J]. 计算机应用, 2020, 40(6): 1799-1805. |
ZHANG J G, LI D P, YANG X D, et al. Deformable medical image registration algorithm based on deep convolution feature optical flow [J]. Journal of Computer Applications, 2020, 40(6): 1799-1805. | |
17 | 王丽芳, 王雁丽, 蔺素珍, 等. 基于改进的Zernike矩的局部描述符与图割离散优化的非刚性多模态脑部图像配准[J]. 计算机应用, 2019, 39(2): 582-588. |
WANG L F, WANG Y L, LIN S Z, et al. Non-rigid multi-modal brain image registration by using improved Zernike moment based local descriptor and graph cuts discrete optimization [J]. Journal of Computer Applications, 2019, 39(2): 582-588. | |
18 | CHEN J, FREY E C, HE Y, et al. TransMorph: Transformer for unsupervised medical image registration [J]. Medical Image Analysis, 2022, 82: 102615. |
19 | QIN Z, YU H, WANG C, et al. Geometric Transformer for fast and robust point cloud registration [C]// Proceedings of the 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2022: 11133-11142. |
20 | MOK T C W, CHUNG A C S. Affine medical image registration with coarse-to-fine vision transformer [C]// Proceedings of the 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2022: 20803-20812. |
21 | MENG C, LI Y, XU Y, et al. A weakly supervised framework for 2D/3D vascular registration oriented to incomplete 2D blood vessels[J]. IEEE Transactions on Medical Robotics and Bionics, 2022, 4(2): 381-390. |
22 | GAO C, KILLEEN B D, HU Y, et al. Synthetic data accelerates the development of generalizable learning-based algorithms for X-ray image analysis [J]. Nature Machine Intelligence, 2023, 5(3): 294-308. |
23 | ZHENG S, YANG X, WANG Y, et al. Unsupervised cross-modality domain adaptation network for X-ray to CT registration[J]. IEEE Journal of Biomedical and Health Informatics, 2022, 26(6): 2637-2647. |
24 | LIAO H, LIN W-A, ZHANG J, et al. Multiview 2D/3D rigid registration via a point-of-interest network for tracking and triangulation [C]// Proceedings of the 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2019: 12630-12639. |
25 | LI P, PEI Y, GUO Y, et al. Non-rigid 2D-3D registration using convolutional autoencoders [C]// Proceedings of the 2020 IEEE 17th International Symposium on Biomedical Imaging. Piscataway: IEEE, 2020: 700-704. |
26 | ZHANG Y. An unsupervised 2D-3D deformable registration network (2D3D-RegNet) for cone-beam CT estimation [J]. Physics in Medicine and Biology, 2021, 66(7): 074001. |
27 | HE K, ZHANG X, REN S, et al. Deep residual learning for image recognition [C]// Proceedings of the 2016 IEEE Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2016: 770-778. |
28 | ZHU J-Y, PARK T, ISOLA P, et al. Unpaired image-to-image translation using cycle-consistent adversarial networks [C]// Proceedings of the 2017 IEEE International Conference on Computer Vision. Piscataway: IEEE, 2017: 2242-2251. |
29 | PASZKE A, GROSS S, MASSA F, et al. PyTorch: an imperative style, high-performance deep learning library [C]// Proceedings of the 33rd International Conference on Neural Information Processing Systems. Red Hook: Curran Associates Inc., 2019: 8026-8037. |
30 | YONG H, HUANG J, HUA X, et al. Gradient centralization: a new optimization technique for deep neural networks [C]// Proceedings of the 16th European Conference on Computer Vision. Cham: Springer, 2020: 635-652. |
31 | LÖFFLER M T, SEKUBOYINA A, JACOB A, et al. A vertebral segmentation dataset with fracture grading [J]. Radiology: Artificial Intelligence, 2020, 2(4): e190138. |
32 | FITZPATRICK J M, WEST J B. The distribution of target registration error in rigid-body point-based registration [J]. IEEE Transactions on Medical Imaging, 2001, 20(9): 917-927. |
33 | YUAN L, CHEN Y, WANG T, et al. Tokens-to-Token ViT: training vision Transformers from scratch on ImageNet [C]// Proceedings of the 2021 IEEE/CVF International Conference on Computer Vision. Piscataway: IEEE, 2021: 538-547. |
34 | DE SILVA T, UNERI A, KETCHA M D, et al. 3D-2D image registration for target localization in spine surgery: investigation of similarity metrics providing robustness to content mismatch [J]. Physics in Medicine and Biology, 2016, 61(8): 3009-3025. |
[1] | 黄云川, 江永全, 黄骏涛, 杨燕. 基于元图同构网络的分子毒性预测[J]. 《计算机应用》唯一官方网站, 2024, 44(9): 2964-2969. |
[2] | 潘烨新, 杨哲. 基于多级特征双向融合的小目标检测优化模型[J]. 《计算机应用》唯一官方网站, 2024, 44(9): 2871-2877. |
[3] | 李顺勇, 李师毅, 胥瑞, 赵兴旺. 基于自注意力融合的不完整多视图聚类算法[J]. 《计算机应用》唯一官方网站, 2024, 44(9): 2696-2703. |
[4] | 秦璟, 秦志光, 李发礼, 彭悦恒. 基于概率稀疏自注意力神经网络的重性抑郁疾患诊断[J]. 《计算机应用》唯一官方网站, 2024, 44(9): 2970-2974. |
[5] | 李力铤, 华蓓, 贺若舟, 徐况. 基于解耦注意力机制的多变量时序预测模型[J]. 《计算机应用》唯一官方网站, 2024, 44(9): 2732-2738. |
[6] | 刘禹含, 吉根林, 张红苹. 基于骨架图与混合注意力的视频行人异常检测方法[J]. 《计算机应用》唯一官方网站, 2024, 44(8): 2551-2557. |
[7] | 顾焰杰, 张英俊, 刘晓倩, 周围, 孙威. 基于时空多图融合的交通流量预测[J]. 《计算机应用》唯一官方网站, 2024, 44(8): 2618-2625. |
[8] | 杨帆, 邹窈, 朱明志, 马振伟, 程大伟, 蒋昌俊. 基于图注意力Transformer神经网络的信用卡欺诈检测模型[J]. 《计算机应用》唯一官方网站, 2024, 44(8): 2634-2642. |
[9] | 石乾宏, 杨燕, 江永全, 欧阳小草, 范武波, 陈强, 姜涛, 李媛. 面向空气质量预测的多粒度突变拟合网络[J]. 《计算机应用》唯一官方网站, 2024, 44(8): 2643-2650. |
[10] | 赵亦群, 张志禹, 董雪. 基于密集残差物理信息神经网络的各向异性旅行时计算方法[J]. 《计算机应用》唯一官方网站, 2024, 44(7): 2310-2318. |
[11] | 徐松, 张文博, 王一帆. 基于时空信息的轻量视频显著性目标检测网络[J]. 《计算机应用》唯一官方网站, 2024, 44(7): 2192-2199. |
[12] | 孙逊, 冯睿锋, 陈彦如. 基于深度与实例分割融合的单目3D目标检测方法[J]. 《计算机应用》唯一官方网站, 2024, 44(7): 2208-2215. |
[13] | 吴筝, 程志友, 汪真天, 汪传建, 王胜, 许辉. 基于深度学习的患者麻醉复苏过程中的头部运动幅度分类方法[J]. 《计算机应用》唯一官方网站, 2024, 44(7): 2258-2263. |
[14] | 李欢欢, 黄添强, 丁雪梅, 罗海峰, 黄丽清. 基于多尺度时空图卷积网络的交通出行需求预测[J]. 《计算机应用》唯一官方网站, 2024, 44(7): 2065-2072. |
[15] | 张郅, 李欣, 叶乃夫, 胡凯茜. 基于暗知识保护的模型窃取防御技术DKP[J]. 《计算机应用》唯一官方网站, 2024, 44(7): 2080-2086. |
阅读次数 | ||||||
全文 |
|
|||||
摘要 |
|
|||||