Algorithm for Automatic Brain-Shift Detection Using the Distance Between Feature Descriptors
Algorithm for Automatic Brain-Shift Detection Using the Distance Between Feature Descriptors
复制标题
使用特征描述符之间的距离进行自动脑转移检测的算法
DOI：
10.1007/978-3-031-05409-9_29
复制
发表时间：
2022
期刊：
影响因子：
--
通讯作者：
Noborio Hiroshi
中科院分区：
文献类型：
--
作者：
Mori Takumi;Nonaka Masahiro;Kunii Takahiro;Koeda Masanao;Watanabe Kaoru;Noborio Hiroshi
在这项研究中，我们进行了基础研究，使用术前和术后标准化（位置，姿势和比例对齐）患者数字成像和医学通信来准确检测术中脑移位，旨在开发一种可与术中传感器basting沿着在线使用的脑模型。为了准确的脑转移检测，我们评估了几种特征点匹配算法检测到的局部脑转移及其参数与人工产生的局部脑转移之间的一致性。结果表明，尺度不变特征变换算法不适合检测脑移位，而加速KAZE算法取得了良好的效果。因此，本研究确定了一个合适的特征点检测算法及其参数的检测大脑的变化。
In this study, we conducted basic research to accurately detect intraoperative brain shifts using preoperative and postoperative normalized (position, posture, and scale aligned) patient Digital Imaging and Communications in Medicine and aimed to develop a brain model that could be used online along with intraoperative sensor basting. For accurate brain-shift detections, we evaluated the agreement between the local brain shifts detected by several feature-point matching algorithms and their parameters and the artificially produced local brain shifts. The results indicated that the scale-invariant feature transform algorithm proved unsuitable for detecting brain shifts, while the accelerated-KAZE algorithm produced good results. Thus, this study identified a suitable feature-point detection algorithm and its parameters for the detection of brain shifts.