IEEE Trans Vis Comput Graph. 2024 Jul;30(7):4362-4374. doi: 10.1109/TVCG.2023.3260001. Epub 2024 Jun 27.
During traditional surgeries, planning and instrument guidance is displayed on an external screen. Recent developments of augmented reality (AR) techniques can overcome obstacles including hand-eye discoordination and heavy mental load. Among these AR technologies, optical see-through (OST) schemes with stereoscopic displays can provide depth perception and retain the physical scene for safety considerations. However, limitations still exist in certain AR systems and the influence of these factors on surgical performance is yet to explore. To this end, experiments of multi-scale surgical tasks were carried out to compare head-mounted display (HMD) AR and autostereoscopic image overlay (AIO) AR, concerning objective performance and subjective evaluation. To solely analyze effects brought by display techniques, the tracking system in each included display system was identical and similar tracking accuracy was proved by a preliminary experiment. Focus and context rendering was utilized to enhance in-situ visualization for surgical guidance. Latency values of all display systems were assessed and a delay experiment proved the latency differences had no significant impact on user performance. Results of multi-scale surgical tasks showed that HMD outperformed in detailed operations probably due to stable resolution along the depth axis, while AIO had better performance in larger-scale operations for better depth perception. This article helps point out the critical limitations of current OST AR techniques and potentially promotes the progress of AR applications in surgical guidance.
在传统手术中,规划和器械引导显示在外部屏幕上。最近,增强现实(AR)技术的发展可以克服手眼不协调和精神负担过重等障碍。在这些 AR 技术中,具有立体显示的光学穿透(OST)方案可以提供深度感知并保留物理场景以确保安全。然而,某些 AR 系统仍然存在局限性,这些因素对手术性能的影响仍有待探索。为此,进行了多尺度手术任务实验,以比较头戴式显示器(HMD)AR 和自动立体图像叠加(AIO)AR,涉及客观性能和主观评估。为了单独分析显示技术带来的影响,每个显示系统中的跟踪系统都是相同的,并且通过初步实验证明了相似的跟踪精度。聚焦和上下文渲染用于增强手术引导的现场可视化。评估了所有显示系统的延迟值,延迟实验证明延迟差异对用户性能没有显著影响。多尺度手术任务的结果表明,HMD 在细节操作中表现更好,可能是因为其在深度轴上具有稳定的分辨率,而 AIO 在较大规模的操作中表现更好,因为其具有更好的深度感知。本文有助于指出当前 OST AR 技术的关键局限性,并有可能促进 AR 在手术引导中的应用进展。