当前位置: X-MOL 学术Neurosurg. Focus › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
Development of an inside-out augmented reality technique for neurosurgical navigation
Neurosurgical Focus ( IF 4.1 ) Pub Date : 2021-08-01 , DOI: 10.3171/2021.5.focus21184
Yun-Sik Dho 1 , Sang Joon Park 2 , Haneul Choi 2 , Youngdeok Kim 2 , Hyeong Cheol Moon 1 , Kyung Min Kim 3 , Ho Kang 3 , Eun Jung Lee 3 , Min-Sung Kim 3 , Jin Wook Kim 3 , Yong Hwy Kim 3 , Young Gyu Kim 1 , Chul-Kee Park 3
Affiliation  

OBJECTIVE

With the advancement of 3D modeling techniques and visualization devices, augmented reality (AR)–based navigation (AR navigation) is being developed actively. The authors developed a pilot model of their newly developed inside-out tracking AR navigation system.

METHODS

The inside-out AR navigation technique was developed based on the visual inertial odometry (VIO) algorithm. The Quick Response (QR) marker was created and used for the image feature–detection algorithm. Inside-out AR navigation works through the steps of visualization device recognition, marker recognition, AR implementation, and registration within the running environment. A virtual 3D patient model for AR rendering and a 3D-printed patient model for validating registration accuracy were created. Inside-out tracking was used for the registration. The registration accuracy was validated by using intuitive, visualization, and quantitative methods for identifying coordinates by matching errors. Fine-tuning and opacity-adjustment functions were developed.

RESULTS

ARKit-based inside-out AR navigation was developed. The fiducial marker of the AR model and those of the 3D-printed patient model were correctly overlapped at all locations without errors. The tumor and anatomical structures of AR navigation and the tumors and structures placed in the intracranial space of the 3D-printed patient model precisely overlapped. The registration accuracy was quantified using coordinates, and the average moving errors of the x-axis and y-axis were 0.52 ± 0.35 and 0.05 ± 0.16 mm, respectively. The gradients from the x-axis and y-axis were 0.35° and 1.02°, respectively. Application of the fine-tuning and opacity-adjustment functions was proven by the videos.

CONCLUSIONS

The authors developed a novel inside-out tracking–based AR navigation system and validated its registration accuracy. This technical system could be applied in the novel navigation system for patient-specific neurosurgery.



中文翻译:

用于神经外科导航的由内而外的增强现实技术的开发

客观的

随着 3D 建模技术和可视化设备的进步,基于增强现实 (AR) 的导航 (AR navigation) 正在积极发展。作者开发了他们新开发的由内而外跟踪 AR 导航系统的试点模型。

方法

由内而外的 AR 导航技术是基于视觉惯性里程计 (VIO) 算法开发的。快速响应 (QR) 标记被创建并用于图像特征检测算法。由内而外的 AR 导航通过可视化设备识别、标记识别、AR 实现和在运行环境中注册的步骤工作。创建了用于 AR 渲染的虚拟 3D 患者模型和用于验证配准准确性的 3D 打印患者模型。注册使用了由内而外的跟踪。通过使用直观、可视化和定量的方法通过匹配误差识别坐标来验证配准准确性。开发了微调和不透明度调整功能。

结果

开发了基于 ARKit 的由内而外的 AR 导航。AR 模型的基准标记和 3D 打印患者模型的基准标记在所有位置正确重叠,没有错误。AR导航的肿瘤和解剖结构与放置在3D打印患者模型颅内空间的肿瘤和结构精确重叠。使用坐标量化配准精度,x轴和y轴的平均移动误差分别为0.52±0.35和0.05±0.16mm。x 轴和 y 轴的梯度分别为 0.35° 和 1.02°。视频证明了微调和不透明度调整功能的应用。

结论

作者开发了一种新颖的基于内向外跟踪的 AR 导航系统并验证了其配准准确性。该技术系统可应用于针对特定患者的神经外科手术的新型导航系统。

更新日期:2021-08-03
down
wechat
bug