• 文献检索
  • 文档翻译
  • 深度研究
  • 学术资讯
  • Suppr Zotero 插件Zotero 插件
  • 邀请有礼
  • 套餐&价格
  • 历史记录
应用&插件
Suppr Zotero 插件Zotero 插件浏览器插件Mac 客户端Windows 客户端微信小程序
定价
高级版会员购买积分包购买API积分包
服务
文献检索文档翻译深度研究API 文档MCP 服务
关于我们
关于 Suppr公司介绍联系我们用户协议隐私条款
关注我们

Suppr 超能文献

核心技术专利：CN118964589B侵权必究
粤ICP备2023148730 号-1Suppr @ 2026

文献检索

告别复杂PubMed语法，用中文像聊天一样搜索，搜遍4000万医学文献。AI智能推荐，让科研检索更轻松。

立即免费搜索

文件翻译

保留排版，准确专业，支持PDF/Word/PPT等文件格式，支持 12+语言互译。

免费翻译文档

深度研究

AI帮你快速写综述，25分钟生成高质量综述，智能提取关键信息，辅助科研写作。

立即免费体验

基于实时呼吸信号预测的长短期记忆方法初始学习算法比较

Comparison of initial learning algorithms for long short-term memory method on real-time respiratory signal prediction.

作者信息

Sun Wenzheng, Dang Jun, Zhang Lei, Wei Qichun

机构信息

Department of Radiation Oncology, The Second Affiliated Hospital, School of Medicine, Zhejiang University, Hangzhou, Zhejiang, China.

Department of Radiation Oncology, National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital and Shenzhen Hospital, Chinese Academy of Medical Science and Peking Union Medical College, Shenzhen, Guangdong, China.

出版信息

Front Oncol. 2023 Jan 20;13:1101225. doi: 10.3389/fonc.2023.1101225. eCollection 2023.

DOI:10.3389/fonc.2023.1101225
PMID:36910606
原文链接:https://pmc.ncbi.nlm.nih.gov/articles/PMC9999041/
Abstract

AIM

This study aimed to examine the effect of the weight initializers on the respiratory signal prediction performance using the long short-term memory (LSTM) model.

METHODS

Respiratory signals collected with the CyberKnife Synchrony device during 304 breathing motion traces were used in this study. The effectiveness of four weight initializers (Glorot, He, Orthogonal, and Narrow-normal) on the prediction performance of the LSTM model was investigated. The prediction performance was evaluated by the normalized root mean square error (NRMSE) between the ground truth and predicted respiratory signal.

RESULTS

Among the four initializers, the He initializer showed the best performance. The mean NRMSE with 385-ms ahead time using the He initializer was superior by 7.5%, 8.3%, and 11.3% as compared to that using the Glorot, Orthogonal, and Narrow-normal initializer, respectively. The confidence interval of NRMSE using Glorot, He, Orthogonal, and Narrow-normal initializer were [0.099, 0.175], [0.097, 0.147], [0.101, 0.176], and [0.107, 0.178], respectively.

CONCLUSIONS

The experiment results in this study indicated that He could be a valuable initializer in the LSTM model for the respiratory signal prediction.

摘要

目的

本研究旨在探讨权重初始化方法对使用长短期记忆（LSTM）模型的呼吸信号预测性能的影响。

方法

本研究使用在304次呼吸运动轨迹期间通过射波刀同步装置收集的呼吸信号。研究了四种权重初始化方法（Glorot、He、正交和窄正态）对LSTM模型预测性能的有效性。通过真实值与预测呼吸信号之间的归一化均方根误差（NRMSE）来评估预测性能。

结果

在这四种初始化方法中，He初始化方法表现出最佳性能。使用He初始化方法提前385毫秒时的平均NRMSE分别比使用Glorot、正交和窄正态初始化方法时高出7.5%、8.3%和11.3%。使用Glorot、He、正交和窄正态初始化方法时NRMSE的置信区间分别为[0.099, 0.175]、[0.097, 0.147]、[0.101, 0.176]和[0.107, 0.178]。

结论

本研究的实验结果表明，He可能是LSTM模型中用于呼吸信号预测的一种有价值的初始化方法。

相似文献

1
Comparison of initial learning algorithms for long short-term memory method on real-time respiratory signal prediction.基于实时呼吸信号预测的长短期记忆方法初始学习算法比较
Front Oncol. 2023 Jan 20;13:1101225. doi: 10.3389/fonc.2023.1101225. eCollection 2023.
2
Real-time liver tracking algorithm based on LSTM and SVR networks for use in surface-guided radiation therapy.基于 LSTM 和 SVR 网络的实时肝脏跟踪算法，用于表面引导放射治疗。
Radiat Oncol. 2021 Jan 14;16(1):13. doi: 10.1186/s13014-020-01729-7.
3
Real-Time Respiratory Tumor Motion Prediction Based on a Temporal Convolutional Neural Network: Prediction Model Development Study.基于时间卷积神经网络的实时呼吸肿瘤运动预测：预测模型开发研究。
J Med Internet Res. 2021 Aug 27;23(8):e27235. doi: 10.2196/27235.
4
Clinical applicability of deep learning-based respiratory signal prediction models for four-dimensional radiation therapy.基于深度学习的呼吸信号预测模型在四维放射治疗中的临床适用性。
PLoS One. 2022 Oct 18;17(10):e0275719. doi: 10.1371/journal.pone.0275719. eCollection 2022.
5
Enhancing PM Prediction Using NARX-Based Combined CNN and LSTM Hybrid Model.基于 NARX 的卷积神经网络和长短期记忆网络混合模型增强 PM 预测。
Sensors (Basel). 2022 Jun 11;22(12):4418. doi: 10.3390/s22124418.
6
Offline and online LSTM networks for respiratory motion prediction in MR-guided radiotherapy.离线和在线长短期记忆网络在磁共振引导放射治疗中的呼吸运动预测。
Phys Med Biol. 2022 Apr 19;67(9). doi: 10.1088/1361-6560/ac60b7.
7
Predicting respiratory motion using a novel patient specific dual deep recurrent neural networks.使用一种新型的患者特异性双深度循环神经网络预测呼吸运动。
Biomed Phys Eng Express. 2022 Sep 29;8(6). doi: 10.1088/2057-1976/ac938f.
8
Cooperative Training of Fast Thinking Initializer and Slow Thinking Solver for Conditional Learning.条件学习中快速思考初始化器和慢速思考求解器的协同训练。
IEEE Trans Pattern Anal Mach Intell. 2022 Aug;44(8):3957-3973. doi: 10.1109/TPAMI.2021.3069023. Epub 2022 Jul 1.
9
Smoothing of respiratory motion traces for motion-compensated radiotherapy.呼吸运动轨迹平滑处理在运动补偿放射治疗中的应用。
Med Phys. 2010 Jan;37(1):282-94. doi: 10.1118/1.3271684.
10
Target localization during respiration motion based on LSTM: A pilot study on robotic puncture system.基于 LSTM 的呼吸运动目标定位：机器人穿刺系统的初步研究。
Int J Med Robot. 2021 Jun;17(3):e2247. doi: 10.1002/rcs.2247. Epub 2021 Mar 4.

本文引用的文献

1
Real-Time Respiratory Tumor Motion Prediction Based on a Temporal Convolutional Neural Network: Prediction Model Development Study.基于时间卷积神经网络的实时呼吸肿瘤运动预测：预测模型开发研究。
J Med Internet Res. 2021 Aug 27;23(8):e27235. doi: 10.2196/27235.
2
Real-time liver tracking algorithm based on LSTM and SVR networks for use in surface-guided radiation therapy.基于 LSTM 和 SVR 网络的实时肝脏跟踪算法，用于表面引导放射治疗。
Radiat Oncol. 2021 Jan 14;16(1):13. doi: 10.1186/s13014-020-01729-7.
3
Adaptive respiratory signal prediction using dual multi-layer perceptron neural networks.
使用双层多层感知机神经网络进行自适应呼吸信号预测。
Phys Med Biol. 2020 Sep 14;65(18):185005. doi: 10.1088/1361-6560/abb170.
4
Towards real-time respiratory motion prediction based on long short-term memory neural networks.基于长短时记忆神经网络的实时呼吸运动预测。
Phys Med Biol. 2019 Apr 10;64(8):085010. doi: 10.1088/1361-6560/ab13fa.
5
Respiratory signal prediction based on adaptive boosting and multi-layer perceptron neural network.基于自适应增强和多层感知器神经网络的呼吸信号预测
Phys Med Biol. 2017 Aug 3;62(17):6822-6835. doi: 10.1088/1361-6560/aa7cd4.
6
Real-time prediction and gating of respiratory motion using an extended Kalman filter and Gaussian process regression.使用扩展卡尔曼滤波器和高斯过程回归对呼吸运动进行实时预测和门控。
Phys Med Biol. 2015 Jan 7;60(1):233-52. doi: 10.1088/0031-9155/60/1/233. Epub 2014 Dec 9.
7
An adaptive fuzzy prediction model for real time tumor tracking in radiotherapy via external surrogates.基于外部替代物的放射治疗中实时肿瘤跟踪的自适应模糊预测模型。
J Appl Clin Med Phys. 2013 Jan 7;14(1):4008. doi: 10.1120/jacmp.v14i1.4008.
8
A novel technique for markerless, self-sorted 4D-CBCT: feasibility study.一种新型无标记、自动分类 4D-CBCT 技术：可行性研究。
Med Phys. 2012 Mar;39(3):1442-51. doi: 10.1118/1.3685443.
9
Predicting the outcome of respiratory motion prediction.预测呼吸运动预测的结果。
Med Phys. 2011 Oct;38(10):5569-81. doi: 10.1118/1.3633907.
10
Adaptive prediction of respiratory motion for motion compensation radiotherapy.用于运动补偿放射治疗的呼吸运动自适应预测
Phys Med Biol. 2007 Nov 21;52(22):6651-61. doi: 10.1088/0031-9155/52/22/007. Epub 2007 Oct 26.