《计算机应用》唯一官方网站 ›› 2023, Vol. 43 ›› Issue (2): 589-594.DOI: 10.11772/j.issn.1001-9081.2021122147

• 多媒体计算与计算机仿真 • 上一篇    

基于姿态编码器的2D/3D脊椎医学图像实时配准方法

徐少康1, 张战成1,2(), 姚浩男1, 邹智伟1, 张宝成3   

  1. 1.苏州科技大学 电子与信息工程学院, 江苏 苏州 215009
    2.苏州科技大学 苏州市虚拟现实智能交互及应用技术重点实验室, 江苏 苏州 215009
    3.中部战区总医院 骨科, 武汉 430070
  • 收稿日期:2021-12-22 修回日期:2022-07-19 接受日期:2022-07-20 发布日期:2022-09-23 出版日期:2023-02-10
  • 通讯作者: 张战成
  • 作者简介:徐少康(1997—),男,江苏淮安人,硕士研究生,主要研究方向:计算机视觉、深度学习、医学图像配准
    姚浩男(1996—),男,江苏徐州人,硕士研究生,主要研究方向:计算机视觉、图像表征学习
    邹智伟(1997—),男,江苏苏州人,硕士研究生,主要研究方向:计算机视觉、图像分割
    张宝成(1983—),男,山西平遥人,硕士,主要研究方向:创伤骨科、骨修复材料、人造血管。
  • 基金资助:
    国家自然科学基金资助项目(61772237)

2D/3D spine medical image real-time registration method based on pose encoder

Shaokang XU1, Zhancheng ZHANG1,2(), Haonan YAO1, Zhiwei ZOU1, Baocheng ZHANG3   

  1. 1.School of Electronic and Information Engineering,Suzhou University of Science and Technology,Suzhou Jiangsu 215009,China
    2.Suzhou Key Laboratory of Virtual Reality Intelligent Interaction and Application Technology,Suzhou University of Science and Technology,Suzhou Jiangsu 215009,China
    3.Department of Orthopaedics,General Hospital of Central Theater Command,Wuhan Hubei 430070,China
  • Received:2021-12-22 Revised:2022-07-19 Accepted:2022-07-20 Online:2022-09-23 Published:2023-02-10
  • Contact: Zhancheng ZHANG
  • About author:XU Shaokang, born in 1997, M. S. candidate. His research interests include computer vision, deep learning, medical image registration.
    YAO Haonan, born in 1996, M. S. candidate. His research interests include computer vision, image representation learning.
    ZOU Zhiwei, born in 1997, M. S. candidate. His research interests include computer vision, image segmentation.
    ZHANG Baocheng, born in 1983, M. S. His research interests include traumatology and orthopaedics, bone repair materials, artificial blood vessel.
  • Supported by:
    National Natural Science Foundation of China(61772237)

摘要:

2D/3D医学图像配准是骨科手术三维实时导航中的一项关键技术,然而传统的基于优化迭代的2D/3D配准方法需要经过多次迭代计算,无法满足医生在手术过程中对于实时配准的要求。针对该问题,提出一种基于自编码器的姿态回归网络来通过隐空间解码捕获几何姿态信息,从而快速地回归出术中X射线图像对应的术前脊椎位置的3D姿态,并经过重新投影生成最终的配准图像。通过引入新的损失函数,以“粗细”结合配准的方式对模型进行约束,保证了姿态回归的精确度。在CTSpine1K脊椎数据集中抽取100组CT扫描图像进行10折交叉验证,实验结果表明:所提出的模型所生成的配准结果图像与X射线图像的平均绝对误差(MAE)为0.04,平均目标配准误差(mTRE)为1.16 mm,单帧耗时1.7 s。与基于传统优化的方法相比,该模型配准时间大幅缩短。相较于基于学习的方法,该模型在快速配准的同时,保证了较高的配准精度。可见,所提模型可以满足术中实时高精配准的要求。

关键词: 2D/3D图像配准, 自编码器, 隐空间, 姿态回归, 骨科手术

Abstract:

2D/3D medical image registration is a key technology in 3D real-time navigation of orthopedic surgery. However, the traditional 2D/3D registration methods based on optimization iteration require multiple iterative calculations, which cannot meet the requirements of doctors for real-time registration during surgery. To solve this problem, a pose regression network based on autoencoder was proposed. In this network, the geometric pose information was captured through hidden space decoding, thereby quickly regressing the 3D pose of preoperative spine pose corresponding to the intraoperative X-ray image, and the final registration image was generated through reprojection. By introducing new loss functions, the model was constrained by “Rough to Fine” combined registration method to ensure the accuracy of pose regression. In CTSpine1K spine dataset, 100 CT scan image sets were extracted for 10-fold cross-validation. Experimental results show that the registration result image generated by the proposed model has the Mean Absolute Error (MAE) with the X-ray image of 0.04, the mean Target Registration Error (mTRE) with the X-ray image of 1.16 mm, and the single frame consumption time of 1.7 s. Compared to the traditional optimization based method, the proposed model has registration time greatly shortened. Compared with the learning-based method, this model ensures a high registration accuracy with quick registration. Therefore, the proposed model can meet the requirements of intraoperative real-time high-precision registration.

Key words: 2D/3D image registration, autoencoder, latent space, pose regression, orthopedic surgery

中图分类号: