摘要
为了解决脊柱手术机器人辅助实施椎弓根植钉过程中的图像辅助定位问题,该文提出了一种基于机器学习策略的椎弓根植钉规划方法。该方法利用卷积神经网络对脊柱计算机断层扫描(CT)图像进行学习和训练,通过建立神经网络模型确定网络内各层间的调整参数,然后对样本图像进行特征提取并分类,采用交叉验证法对样本数据进行训练,验证卷积神经网络模型的正确性。通过机器学习方法对计算机断层扫描图像中适合做椎弓根植钉手术规划的图像区域进行识别,从而快速定位到植钉安全约束区域,并通过相应的图像处理方法实现植钉操作规划。医生只需要基于安全约束区域内的植钉规划完成最终的手术任务规划,能够显著提升手术效率。
In the process of pedicle screw insertion surgery assisted by spinal surgery robot,the surgeon needs to spend a lot of time in searching and determining the location suitable for pedicle screw operation.In order to improve the operation efficiency,a rapid method based on machine learning was investigated to realize efficient planning of robot-assisted pedicle screw insertion.In the proposed method,the convolution neural network that commonly used in artificial intelligence research was introduced based on training set of spine computed tomography images.Firstly,the neural network model was set up to determine the adjustment parameters between layers in the network.Then the feature extraction and classification of sample image were carried out.The cross validation method was used to train sample data to verify the correctness of convolution neural network model.Finally,the method of machine learning was used to identify the computed tomography images of patients that suitable for the location of pedicle screw implant.With this proposed method,surgeons only need to complete the final fine planning based on the area of safety constraint,and efficiency of operation can be improved distinctly.
作者
苏全健
孙宇
齐晓志
SU Quanjian1,2 SUN Yu1,2 QI Xiaozhi1(1 Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, Shenzhen 518055, China ) 2( Shenzhen Graduate School, Harbin Institutes of Technology, Shenzhen 518055, China)
出处
《集成技术》
2018年第3期42-53,共12页
Journal of Integration Technology
基金
国家重点研发计划项目(2017YFC0110600)
国家自然科学基金联合基金项目(U1713221)
国家自然科学基金面上项目(61573336)
国家自然科学基金青年项目(51705512)
深圳市基础研究(学科布局)项目(JCYJ20160608153218487
JCYJ20170413104438332)
关键词
椎弓根植钉规划
机器学习
图像分类
图像识别
卷积神经网络
planning of pedicle screw insertion
machine learning
image classification
image identify
convolution neural network