CV-Based Analysis for Microscopic Gauze Suturing Training
CV-Based Analysis for Microscopic Gauze Suturing Training
复制标题
DOI：
10.1145/3458709.3458991
复制
发表时间：
2021-02
期刊：
影响因子：
--
通讯作者：
Mikihito Matsuura;Shio Miyafuji;Erwin Wu;S. Kiyofuji;Taichi Kin;T. Igarashi;H. Koike
中科院分区：
文献类型：
--
作者：
Mikihito Matsuura;Shio Miyafuji;Erwin Wu;S. Kiyofuji;Taichi Kin;T. Igarashi;H. Koike
本文提出了一个显微缝合实践支持系统的基础，旨在减少神经外科医生练习显微纱布缝合所需的时间。该系统从显微镜摄像头的视频中实时检测仪器并提供即时分析。当学员完成练习后，他们可以立即查看他们的结果。我们介绍了一个连续的图像数据集，其中的手术阶段，以及手术器械的边界框，注释。YOLO V4网络使用建议的数据集进行了微调，并实现了约94%的准确度。我们还提出了一种工具，用于在每次训练后使用基于动态规划的算法从跟踪数据进行相位估计，这使我们能够以约83%的准确度估计先前练习会话的相位。这用于应用程序检测关注点并提供反馈。这一建议增强了从业者的技能获取，使他们能够立即反映他们最近的实践会议，而不是漫无目的地重复。
This paper proposes the basis of a microscopic suture practice support system aimed to reduce the time required for neurosurgeons to practice microscopic gauze suturing. The system detects instruments in real-time from the video of the microscope camera and provides an immediate analysis. After practitioners have completed practicing, they can immediately view their results. We introduce a sequential image dataset in which the surgery phases, as well as the bounding boxes of surgical instruments, are annotated. A YOLO V4 network is fine-tuned with the proposed dataset and achieves an accuracy of approximately 94%. We also propose a tool for a phase estimation after each suturing using a Dynamic Programming based algorithm from the tracking data, which allows us to estimate the phase of the previous practice session with about 83 % accuracy. This is used for an application to detect and provide feedback on points of concern. This proposal augments the practitioners’ acquisition of skills, allowing them to immediately reflect on their most recent practice session, rather than repeating aimlessly.