基于记忆学习法的放疗中呼吸运动预测技术的研究

doi:10.3969/j.issn.0258-8021. 2014. 02.003

摘要
图/表
参考文献
相关文章 (10)

全文: PDF (1871 KB) HTML (1 KB)
输出: BibTeX | EndNote (RIS)

摘要对胸腹部肿瘤进行实时跟踪放疗时，需要通过预测来补偿系统延迟。然而，由于呼吸运动的复杂性，传统方法难以满足要求。本文应用一种基于记忆学习法进行呼吸预测，该方法首先存储训练数据到记忆中，然后查找相关数据应答当前查询。在此基础上，采用“滑窗法”动态更新训练数据集，并针对预测过程中出现的 “病态矩阵”采用脊回归进一步改进算法，使算法的精确性和鲁棒性有了很大提高。实验使用POLARIS红外定位系统采集了10例正常人体表的红外反射标记物的呼吸运动数据样本，平均幅度约为20 mm(9.2～37.8 mm)，采用改进后的基于记忆学习法（预测步长为1 s），平均绝对误差约为0.3 mm(0.08～0.8 mm)，每次估值耗时约1 ms。所提出的方法能够准确和实时捕捉复杂的呼吸运动轨迹。

	服务

	把本文推荐给朋友
	加入我的书架
	加入引用管理器
	E-mail Alert
	RSS
	作者相关文章
	万伟权¹张慧连^{2 &}

关键词 ：放射治疗, 记忆学习法, 呼吸运动, 实时, 预测

Abstract：Prediction is necessary to compensate the system latency in the realtime tracking radiation therapy for thoracic and abdominal cancers. However, because of the complexity of the breathing motion, conventional methods are far from clinical requirements. This paper proposed a memorybased learning method to predict respiratory motion. The method stores the training data in memory, then finds relevant data to answer a particular query. Furthermore, the paper adopts dynamic update the training data method and ridge regression aimed at “illcondition matrix” to greatly improve the accuracy and robustness of the algorithm. Our experiment collected ten respiratory motion data with average amplitude of 20 mm (9.2～37.8 mm) from humans’ body surface using POLARIS infrared positioning system. Using our methods (prediction horizon is 1s), mean absolute error (MAE) was reduced to 0.3 mm (0.08～0.8 mm), per estimate takes 1 ms. The results confirm that the proposed method is able to capture highly complex breathing movement accurately in real time.

Key words： radiotherapy memorybased learning respiratory motion realtime prediction

基金资助:广东省重大科技专项 (2012A080104010)

引用本文:

万伟权¹张慧连^{2 &. 基于记忆学习法的放疗中呼吸运动预测技术的研究[J]. 中国生物医学工程学报, 2014, 33(2): 148-154.

WAN WeiQuan¹ZHANG HuiLian²XU ZiHai³HE ZhiQiang^{1 . The Study on Predicting Respiratory Motion via MemoryBased Learning in Radiotherapy. journal1, 2014, 33(2): 148-154.}}

链接本文:

http://cjbme.csbme.org/CN/10.3969/j.issn.0258-8021. 2014. 02.003 或 http://cjbme.csbme.org/CN/Y2014/V33/I2/148