北京生物医学工程

一种基于活动形状模型算法的耳郭穴区定位方法

A location method of auricular point area based on active shape model algorithm

作者：王逸卉江淼黄娜温军玲

单位：上海中医药大学针灸推拿学院（上海201203）

关键词：中医；耳穴疗法；耳穴分区；图像处理；活动形状模型

分类号：R318

出版年·卷·期（页码）：2021·40·2（145-150）

摘要：

目的耳穴疗法是中医的重要组成部分，而耳穴定位是治疗的前提。由于耳郭面积较小、耳穴众多导致自动寻穴十分麻烦，本文提出一种基于图像处理技术的耳郭穴区定位方法。方法通过对来自于USTB人耳图像库和自行采集图像共30幅构建训练集，在最新耳穴国家标准GB/T13734-2008的基础上，选取包含耳郭边缘点、曲率极大值点、等分点等在内的65个特征点构建点分布模型（point distribution model, PDM），采用基于活动形状模型(active shape model, ASM)的图像处理方法，对待测人耳图像进行搜索匹配后，通过分别连接构成耳郭穴区的特征点，实现对人耳图像的穴区定位。最后通过计算定位后与人工标定的特征点之间的欧几里得距离，对定位的精度进行评估。结果该方法能初步实现耳郭的穴区定位，欧几里得平均距离为6.246±0.429；结论基于活动形状模型的耳郭穴区定位方法能初步实现人耳图像的穴区定位，有利于耳穴疗法的穴区定位、耳穴自动化仪器的开发和中医耳穴示教等

Objective Auricular therapy is an important part of traditional Chinese medicine, and the location of auricular point is the premise of treatment. But it is hard to locate the auricular points automatically because of too small auricular area and too many auricle points. This paper proposes a location method of auricular point area based on image processing technology. Methods A training set of 30 human ear images from the Human Ear Recognition Laboratory of University of Science and Technology Beijing was constructed. Then integrated with the latest national standard of auricular points (GB /T13734-2008), 65 feature points including auricle edge point, curvature maximum point and bisection point were selected for each image to establish a point distribution model (PDM). Next an image processing method based on active shape model (ASM) was used to search the object ear image. And then through connecting the feature points of each auricular point area, the location of ear image was realized. Finally, the accuracy was evaluated by using Euclidean distance evaluation index between the location results of auricular points and the location of artificial landmarks. Results Our method can realize the location of auricular point areas. The average Euclidean distance value is 6.246 ± 0.429. Conclusions The auricular points location method based on active shape model can realize the location of auricular point area, which is benefit to the development of automatic instrument for auricular therapy and can help the teaching of the related ear points of traditional Chinese medicine.

参考文献：

[1] 程爵棠. 耳穴疗法治百病[M]. 4版. 郑州：河南科学技术出版社, 2017：1-7.

[2] 刘继洪,许艺燕,徐光镇,等.耳穴医疗是中西医结合的一座“桥梁”[J].中国中西医结合杂志,2019,39(6):750-752.

Liu JH, Xu YY, Xu GZ, et al. Auricular medicine is the "bridge" of the integrated traditional Chinese medicine and western medicine[J]. Chinese Journal of Integrated Traditional and Western Medicine, 2019, 39(6): 750-752.

[3] 陈盼碧,杨孝芳,王兴桂,等.图示教学法在耳针法耳穴定位教学中的运用[J].亚太传统医药, 2019,15(6): 214-215.

[4] Yang YT, Huang QF, Jia YF, et al. Current situation and analysis of clinical application of auricular acupoint sticking therapy[J]. Journal of Acupuncture and Tuina Science, 2016,14(2):141-148.

[5] 中华人民共和国国家质量监督检验检疫总局.GB/T 13734—2008耳穴名称与定位[S].北京:中国标准出版社, 2008.

[6] 李一波,黄增喜,张海军,等.一种基于改进GVF Snake的自动人耳检测方法[J].模式识别与人工智能, 2010, 23(4): 552-559.

Li YB, Huang ZX, Zhang HJ, et al. An automatic ear detection method based on improved GVF Snake[J]. Pattern Recognition and Artificial Intelligence, 2010, 23(4): 552-559.

[7] 李素娟,封筠,牛江川.人耳识别中归一化对特征提取和识别影响的研究[J].电脑知识与技术,2008,4(36):2984-2985.

Li SJ, Feng J, Niu JC. The research of the impact of normalization for feature extraction and recognition on ear biomitrice recognition[J].Computer Knowledge and Technology,2008,4(36): 2984-2985.

[8] 高淑欣,穆志纯.人耳识别中的图像归一化研究[J].控制工程,2008,15(1):18-21.

Gao SX, Mu ZC. On image normalization in ear recognition[J]. Control Engineering of China, 2008,15(1): 18-21.

[9] 杨阳.基于ASM算法的肺部轮廓提取研究[D].沈阳：东北大学, 2015.

Yang Y. A study on ASM-based lung edge extraction[D]. Shenyang: Northeastern University, 2015.

[10] 黄飞,尤启房,杨晋吉. ASM的手骨提取方法研究[J].计算机工程与应用,2016,52(3):164-168,219.

Huang F,You QF,Yang JJ. Hand bone extraction method research based on ASM[J]. Computer Engineering and Applications, 2016,52(3):164-168,219.

[11] 常梦龙.基于特征点定位算法的面部穴位定位方法的研究[D]. 北京:北京工业大学, 2017.

Chang ML. Research on facial acupuncture point positioning method based on feature point location algorithm[D]. Beijing: Beijing University of Technology, 2017.

[12] 易佳,王碧琳,刘姝秀,等.结合ASM算法和边缘检测的人脸定位方法研究[J].信息系统工程, 2015(12):125-127.

[13] Cootes TF, Taylor CJ, Cooper DH, et al. Active shape models-their training and application[J]. Computer Vision and Image Understanding, 1995,61(1): 38-59.

[14] 张学.视频中多姿态人脸的特征点定位研究[D].天津：河北工业大学,2015.

Zhang X. Research on feature points location of multi-pose face in video[D].Tianjin: Hebei University of Technology, 2015.

[15] 刘腾龙.视频人脸替换中目标分割算法的研究[D].天津：河北工业大学, 2018.

Liu TL.Research on target segmentation algorithm in video face replacement[D]. Tianjin: Hebei University of Technology, 2018.

服务与反馈：

【文章下载】【加入收藏】

提示：您还未登录，请登录！点此登录