[1]李宏彬,贺太平.种决策树同源算法在肝部B超计算机辅助诊断中的应用比较[J].医学信息,2021,34(19):13-18.[doi:10.3969/j.issn.1006-1959.2021.19.003]
　LI Hong-bin,HE Tai-ping.Comparison of Three Decision Tree Homology Algorithms in ComputerAided Diagnosis of Liver B Ultrasound[J].Medical Information,2021,34(19):13-18.[doi:10.3969/j.issn.1006-1959.2021.19.003]
点击复制

种决策树同源算法在肝部B超计算机辅助诊断中的应用比较()
分享到：

医学信息[ISSN:1006-1959/CN:61-1278/R]

卷:
34卷
期数:
2021年19期
页码:
13-18
栏目:
出版日期:
2021-10-01

文章信息/Info

Title:
Comparison of Three Decision Tree Homology Algorithms in ComputerAided Diagnosis of Liver B Ultrasound
文章编号:
1006-1959（2021）19-0013-06
作者:
李宏彬1; 贺太平2
1.咸阳职业技术学院医学院，陕西 咸阳 712000；2.陕西中医药大学附属医院影像科，陕西 咸阳 712000
Author(s):
LI Hong-bin1; HE Tai-ping2
1.Medical college,Xianyang Vocational and Technical College,Xianyang 712000,Shaanxi,China;2.Imaging department,Affiliated Hospital of Shaanxi University of Traditional Chinese Medicine,Xianyang 712000,Shaanxi,Chin
关键词:
B超; CART; 随机森林; 极端随机树; 肝脏疾病
Keywords:
B ultrasound; CART; Random forest; Extreme random tree; Liver diseases
分类号:
TP391.41
DOI:
10.3969/j.issn.1006-1959.2021.19.003
文献标志码:
A
摘要:
目的 探索CART、随机森林和极端随机树三种决策树同源算法结合不同的特征集进行肝部B超图像的分类效果，以期改善计算机辅助B超诊断肝脏疾病的准确率。方法 从陕西中医药大学附属医院PACS系统下载包含正常肝、脂肪肝、图像增粗、肝硬化、肝部占位和肝囊肿和非肝组织的B超图像，通过框选B超图像生成感兴趣区图块数据，分别对CART、随机森林和极端随机树算法结合7种不同的特征集使用10折的交叉校验进行训练和测试，比较各算法的准确率。结果 在特定的B超切面，随机森林算法结合C7特征集“直方图256 +Python标准纹理72+图块定位3+补充纹理48”的正确率最高，达96.31%；另外，3种算法结合C7都能获得很好的交叉校验正确率。结论 三种决策树同源算法结合C7特征集在B超计算机辅助诊断肝脏疾病中都有较高的价值，其中随机森林算法的表现最优。
Abstract:
Objective To explore the classification effect of CART, random forest and extreme random tree decision tree homologous algorithm combined with different feature sets for liver B ultrasound images, in order to improve the accuracy of computer aided B ultrasound diagnosis of liver diseases.Methods The B-mode ultrasound images of normal liver, fatty liver, image thickening, cirrhosis, liver occupying, liver cyst and non-liver tissue were downloaded from PACS system of Affiliated Hospital of Shaanxi University of Traditional Chinese Medicine. The block data of interest area were generated by selecting B-mode ultrasound images. The CART, random forest and extreme random tree algorithms were trained and tested with seven different feature sets using tenfold cross-validation, and the accuracy of each algorithm was compared.Results In the specific sections of B ultrasound, random forest combined with C7 features set (Histogram 256 plus Python standard textures 72 plus block location 3 plus supplementary textures 48) had the highest accuracy, up to 96.31%, and these three algorithms combined with C7 all could achieve good cross validation accuracy.Conclusion These three decision tree homology algorithms combined with C7 features set could have good application prospects in computer-aided diagnosis of B ultrasound, and random forest is the best.

参考文献/References:

[1]李海强.肝脏超声图像的计算机辅助诊断识别研究[D].青岛:青岛大学,2019.[2]孙逸芳,董磊.计算机辅助系统联合超声对甲状腺结节的鉴别诊断价值[J].实用医药杂志,2019,36(1):30-32,36.[3]刘隆忠,李擎,龙杏章,等.基于计算机辅助诊断技术的超声图像处理软件对甲状腺结节诊断效能的初步研究[J].中华医学超声杂志(电子版),2018,15(12):67-72.[4]蒋恒,王磊,夏开建,等.基于灰度共生矩阵的ADC纹理分析鉴别直肠癌T3亚分期的临床价值[J].中国医学计算机成像杂志,2019,25(6):48-52.[5]刘超然,王宁华,李威,等.年龄对肌肉超声图像纹理特征的影响[J].中国康复医学杂志,2020,35(1):33-39. [6]李光,姜春雪,刘争战,等.Laws纹理能量结合灰度共生矩阵的遥感影像面状地物提取[J].测绘与空间地理信息,2017,40(7):179-181.[7]迟殿委.一种改进的决策树ID3算法的应用[J].现代计算机,2019(17):43-45.[8]安葳鹏,尚家泽.决策树C4.5算法的改进与分析[J].计算机工程与应用,2019,55(12):169-173.[9]Tang R,Zhang X.CART Decision Tree Combined with Boruta Feature Selection for Medical Data Classification[C]//2020 5th IEEE International Conference on Big Data Analytics (ICBDA).2020.[10]Ahmad MW,Mourshed M,Rezgui Y.Trees vs Neurons: Comparison between random forest and ANN for high-resolution prediction of building energy consumption[J].Energy and Buildings,2017(147):77-89.[11]全雪峰.基于随机森林的乳腺癌计算机辅助诊断[J].软件,2017,38(3):57-59.[12]李长胜,王瑜,肖洪兵,等.基于随机森林算法的阿尔茨海默症医学影像分类[J].中国医学物理学杂志,2020,37(8):1005-1009.[13]胡会会,龚敬,聂生东.基于集成随机森林模型的肺结节良恶性分类[J].计算机应用研究,2018,35(10):243-246,251.[14]姚冰莹,李超,邹贵红.基于随机森林的宫颈病变识别应用研究[J].电脑与电信,2018(11):15-17.[15]Soltaninejad M,Yang G,Lambrou T,et al.Automated brain tumour detection and segmentation using superpixel-based extremely randomized trees in FLAIR MRI[J].International Journal of Computer Assisted Radiology and Surgery,2017,12(2):183-203.

相似文献/References:

[1]林秋花,郭海珊,唐 莉,等.影像学检查在诊断胎盘植入中的预测价值[J].医学信息,2019,32(03):58.[doi:10.3969/j.issn.1006-1959.2019.03.019]
　LIN Qiu-hua,GUO Hai-shan,TANG Li,et al.Predictive Value of Imaging Examination in the Diagnosis of Placenta Implantation[J].Medical Information,2019,32(19):58.[doi:10.3969/j.issn.1006-1959.2019.03.019]
[2]李剑飞,苏文玲,梁 柱,等.CT与B超引导下经皮肺穿刺诊断肺癌的价值[J].医学信息,2021,34(09):173.[doi:10.3969/j.issn.1006-1959.2021.09.047]
　LI Jian-fei,SU Wen-ling,LIANG Zhu,et al.The Value of CT and B-ultrasound Guided Percutaneous Lung Puncture in the Diagnosis of Lung Cancer[J].Medical Information,2021,34(19):173.[doi:10.3969/j.issn.1006-1959.2021.09.047]
[3]尹慧慧.B超检查在不同类型急性阑尾炎诊断中的应用价值[J].医学信息,2021,34(24):178.[doi:10.3969/j.issn.1006-1959.2021.24.048]
　YIN Hui-hui.Application Value of B-ultrasound Examination in the Diagnosis of Different Types of Acute Appendicitis[J].Medical Information,2021,34(19):178.[doi:10.3969/j.issn.1006-1959.2021.24.048]

更新日期/Last Update: 1900-01-01