腰麻领域基于卷积神经网络的脊髓圆锥末端位置定位算法  

Location Algorithm of Conus Medullaris End Based on Convolution Neural Network for Lumbar Anesthesia

作　　者：罗燕 李欣 秦小勇[3] 梁晶晶 关延顺 杨建威 LUO Yan;LI Xin;QIN Xiaoyong;LIANG Jingjing;GUAN Yanshun;YANG Jianwei(Department of Anesthesiology,Mianyang Traditional Chinese Medicine Hospital,Mianyang Sichuan 621000,China;Mudanjiang Traditional Chinese Medicine Hospital,Mudanjiang Heilongjiang 157000,China;Sichuan Jiuzhou Electrical Appliance Group Co.,Ltd.,Mianyang Sichuan 621000,China;Nangang Branch of Heilongjiang Academy of Traditional Chinese Medicine,Harbin Heilongjiang 150006,China)

机构地区：[1]绵阳市中医医院麻醉科,四川绵阳621000 [2]牡丹江市中医医院,黑龙江牡丹江157000 [3]四川九洲电器集团有限责任公司,四川绵阳621000 [4]黑龙江省中医药科学院南岗分院,黑龙江哈尔滨150006

出　　处：《中国医疗设备》2020年第1期61-63,81,共4页China Medical Devices

基　　金：牡丹江市科学技术计划项目（Z2018s094）

摘　　要：传统的脊髓圆锥末端内麻醉无可视化技术支持,主要通过麻醉师根据患者体表骨性特征进行穿刺定位,其难易程度与患者当前状态、麻醉师经验密切相关。本文提出了一种基于深度学习的识别定位算法,该算法利用深度学习模型对归一化的目标区域进行深度特征提取,增强特征的表征能力;最后输入到本文采用的SVM分类模型,得到最终的检测到的脊髓圆锥末端结果。定性定量实验结果表明,本文所提出的深度检测模型的检测性能较好,具有一定的可行性和使用价值。There is no visual technical support for conventional conus medullaris end anesthesia,mainly through the puncture location of the anesthesiologist according to the osseous characteristics of the patient’s body surface.The degree of difficulty is closely related to the patient’s current state and anesthesiologist’s experience.In this paper,a recognition and localization algorithm based on deep learning was proposed.The algorithm used the deep learning model to extract the features of the normalized object area,and enhanced the representation ability of the features.Finally,the SVM classification model was used to get the final detection result of the conus medullaris end.The results of qualitative and quantitative experiments showed that the performance of the deep learning detection model proposed in this paper was good,and had certain feasibility and application value.

关 键 词：脊髓圆锥末端 深度学习 麻醉 支持矢量机 目标识别 

分 类 号：R318[医药卫生—生物医学工程] R614[医药卫生—基础医学]