肝包虫病和肝囊肿CT图像的分型研究(1)
摘 要:目的 探讨计算机辅助诊断技术在肝包虫病和肝囊肿CT图像分型中的应用。方法 对单囊型肝包虫病和单发性肝囊肿CT图像感兴趣区域,分别使用传统的预处理方法和图像融合方法,提取原始ROI、预处理后的和融合后的ROI图像Haar小波、DB2小波、Tamura、Gabor滤波器和灰度-梯度共生矩阵特征,通过支持向量机和BP神经网络分类模型分类,比较三种方法的分类准确率,并对各分类模型进行参数评估。结果 从原始ROI图像直接提取的Haar小波、DB2小波、Tamura和GGCM特征的最佳分类准确率均达到了95%以上;融合后的ROI图像五种特征的分类准确率都较高,在90%以上。结论 本研究所使用的方法应用于肝包虫病和肝囊肿CT图像的分型中具有一定的分类优势,为影像学诊断提供依据。
关键词:肝包虫病;肝囊肿;图像融合;特征提取;图像分类
中图分类号:R532.32;R575.4 文献标识码:A DOI:10.3969/j.issn.1006-1959.2018.23.018
, http://www.100md.com
文章编号:1006-1959(2018)23-0061-06
Abstract:Objective To discuss the application of computer aided diagnosis in classification of hepatic hydatid disease and hepatic cyst CT images.Methods For the region of the CT image of single cystic hepatic echinococcosis and single hepatic cyst, the original ROI, pre-processed and fused ROI image Haar wavelet, DB2 were extracted using traditional pre-processing methods and image fusion methods, respectively. Wavelet, Tamura, Gabor filter and gray-gradient co-occurrence matrix characteristics are classified by support vector machine and BP neural network classification model. The classification accuracy of the three methods is compared, and the parameters of each classification model are evaluated. Results The best classification accuracy of Haar wavelet, DB2 wavelet, Tamura and GGCM features extracted from the original ROI image reached more than 95%. The classification accuracy of the five characteristics of the ROI image after fusion is higher,above 90%.Conclusion The method used in this study has a certain classification advantage in the classification of CT images of hepatic hydatidosis and hepatic cysts, and provides a basis for imaging diagnosis.
, http://www.100md.com
Key words:Hepatic echinococcosis;Hepatic cyst;Image fusion;Feature extraction; Image classification
肝包蟲病(hepatic echinococcosis)是一种人畜共患的寄生虫病,由棘球绦幼虫侵染人体肝脏所致,又名肝棘球蚴病,从病理上分为肝细粒棘球蚴病(也称肝包虫囊肿)和肝泡状棘球蚴病(也叫肝的泡状包虫病)两种,肝包虫囊肿较多见[1]。肝包虫病的发病具有地域性特点,多发于牧区,在我国畜牧业发达的新疆、西藏、内蒙、宁夏等省区较常见,当地有着“虫癌”之称。肝囊肿是常见的肝脏良性病变,又称非寄生虫性肝囊肿,大部分是一种先天性疾病[2]。肝包虫病和肝囊肿均起病隐匿,潜伏期较长,初期症状不明显,临床表现相似,易误诊[3]。计算机辅助诊断(computer aided diagnosis,CAD)[4,5]技术作为近几年最为热门的辅助诊断技术,能够辅助医生更加有效的进行判断和确诊疾病,可减少或避免疾病的误诊误治。本研究选取单囊型肝包虫病和单发性肝囊肿CT图像,对原始图像感兴趣区域分别使用传统的预处理方法和图像融合方法,对预处理后的,融合后的和没进过任何处理的感兴趣区域图像提取DB2小波、Haar小波、Gabor滤波、Tamura和灰度-梯度共生矩阵特征,通过支持向量机和BP神经网络进行分类,比较三种方法下的五种特征分类准确率,研究适合于肝包虫病和肝囊肿分型的计算机辅助诊断方法,为临床医师提供决策参考,降低误诊率。
1材料与方法
1.1实验对象 本研究随机选取单囊型肝包虫病和单发性肝囊肿CT图像各120幅,共240幅,由新疆医科大学各附属医院影像中心提供。经影像科医师指导分类,人工干预下从原始图像中分割出病灶信息区,即感兴趣区域(region of interest),见图1。, http://www.100md.com(排孜丽耶·尤山塔依 严传波 木拉提·哈米提)
关键词:肝包虫病;肝囊肿;图像融合;特征提取;图像分类
中图分类号:R532.32;R575.4 文献标识码:A DOI:10.3969/j.issn.1006-1959.2018.23.018
, http://www.100md.com
文章编号:1006-1959(2018)23-0061-06
Abstract:Objective To discuss the application of computer aided diagnosis in classification of hepatic hydatid disease and hepatic cyst CT images.Methods For the region of the CT image of single cystic hepatic echinococcosis and single hepatic cyst, the original ROI, pre-processed and fused ROI image Haar wavelet, DB2 were extracted using traditional pre-processing methods and image fusion methods, respectively. Wavelet, Tamura, Gabor filter and gray-gradient co-occurrence matrix characteristics are classified by support vector machine and BP neural network classification model. The classification accuracy of the three methods is compared, and the parameters of each classification model are evaluated. Results The best classification accuracy of Haar wavelet, DB2 wavelet, Tamura and GGCM features extracted from the original ROI image reached more than 95%. The classification accuracy of the five characteristics of the ROI image after fusion is higher,above 90%.Conclusion The method used in this study has a certain classification advantage in the classification of CT images of hepatic hydatidosis and hepatic cysts, and provides a basis for imaging diagnosis.
, http://www.100md.com
Key words:Hepatic echinococcosis;Hepatic cyst;Image fusion;Feature extraction; Image classification
肝包蟲病(hepatic echinococcosis)是一种人畜共患的寄生虫病,由棘球绦幼虫侵染人体肝脏所致,又名肝棘球蚴病,从病理上分为肝细粒棘球蚴病(也称肝包虫囊肿)和肝泡状棘球蚴病(也叫肝的泡状包虫病)两种,肝包虫囊肿较多见[1]。肝包虫病的发病具有地域性特点,多发于牧区,在我国畜牧业发达的新疆、西藏、内蒙、宁夏等省区较常见,当地有着“虫癌”之称。肝囊肿是常见的肝脏良性病变,又称非寄生虫性肝囊肿,大部分是一种先天性疾病[2]。肝包虫病和肝囊肿均起病隐匿,潜伏期较长,初期症状不明显,临床表现相似,易误诊[3]。计算机辅助诊断(computer aided diagnosis,CAD)[4,5]技术作为近几年最为热门的辅助诊断技术,能够辅助医生更加有效的进行判断和确诊疾病,可减少或避免疾病的误诊误治。本研究选取单囊型肝包虫病和单发性肝囊肿CT图像,对原始图像感兴趣区域分别使用传统的预处理方法和图像融合方法,对预处理后的,融合后的和没进过任何处理的感兴趣区域图像提取DB2小波、Haar小波、Gabor滤波、Tamura和灰度-梯度共生矩阵特征,通过支持向量机和BP神经网络进行分类,比较三种方法下的五种特征分类准确率,研究适合于肝包虫病和肝囊肿分型的计算机辅助诊断方法,为临床医师提供决策参考,降低误诊率。
1材料与方法
1.1实验对象 本研究随机选取单囊型肝包虫病和单发性肝囊肿CT图像各120幅,共240幅,由新疆医科大学各附属医院影像中心提供。经影像科医师指导分类,人工干预下从原始图像中分割出病灶信息区,即感兴趣区域(region of interest),见图1。, http://www.100md.com(排孜丽耶·尤山塔依 严传波 木拉提·哈米提)