基于张量法的阿尔兹海默症脑图像分类

发布时间：2018-01-27 10:43

本文关键词： 阿尔兹海默症轻度认知障碍张量认知分数独立成分分析支持张量机递归特征消除　出处：《中山大学学报(自然科学版)》2017年02期 　论文类型：期刊论文

【摘要】：为了识别阿尔兹海默症(Alzheimer's Disease,AD)与轻度认知障碍(Mild Cognitive Impairment,MCI)患者,提出了一种基于三阶张量方法的以MRI图像脑灰质灰度为特征的分类方法。采集了70例AD患者,112例MCI患者(包含在随访中转化为AD的,MCI-C:MCI Converters与未转化为AD的,MCI-NC:MCI Non-converters各56例),以及70例正常人(NC)的MRI脑图像,提取脑灰质各体素的灰度,获得三阶灰度张量。采用基于张量的独立成分分析,以取得三阶灰度张量的独立成分;为了降低特征维数,利用支持张量机,将张量特征转化为向量特征,再利用递归特征消除法获取有效的主要特征。最后,对四组人群进行分类:AD-NC,MCINC,AD-MCI,MCI-C-MCI-NC,此分类模型采用7折交叉验证的方法进行训练测试。此外,还结合样本的基本信息与认知分数进行分类,证明了基本信息、认知分数和脑灰质灰度提供了互补的信息,有助于提升分类效果。结果表明,该方法拥有优良的分类性能,有助于对AD与MCI的诊断治疗。
[Abstract]:To identify Alzheimer's disease (Alzheimer's disease) and mild cognitive impairment (MCI) Mild Cognitive Impairment. A classification method based on third-order Zhang Liang method was proposed for classification of gray matter in MRI images. 70 AD patients were collected. One hundred and twelve patients with MCI (including MCI-C: MCI Converters who were converted to AD during follow-up) and those who did not. MRI brain images of 56 cases of MCI-NC:MCI Non-converters and 70 cases of normal controls were used to extract grayscale of gray matter. The third grayscale Zhang Liang was obtained. The independent component analysis based on Zhang Liang was used to obtain the independent component of the third grayscale Zhang Liang. In order to reduce the feature dimension, we use Zhang Liang machine to transform Zhang Liang features into vector features, and then use recursive feature elimination method to obtain effective main features. Finally, four groups of people are classified into four groups: AD-NC. MCINCU AD-MCIU MCI-C-MCI-NC.The classification model was tested by a 7-fold cross-validation method. Combining the basic information of the sample with the cognitive score, it is proved that the basic information, the cognitive score and the gray level of gray matter provide complementary information, which is helpful to improve the classification effect. This method has good classification performance and is helpful to the diagnosis and treatment of AD and MCI.
【作者单位】：南方医科大学生物医学工程学院//广东省医学图像处理重点实验室;华南理工大学数学学院;
【基金】：国家自然科学基金(11201219) 广东省医学图像处理重点实验室项目(2014B030301042)
【分类号】：TP391.41;R749.16
【正文快照】： 阿尔兹海默症(AD)是一种起病隐匿的进行性发展的神经系统退行性疾病,其多发于65岁以上的人群。2015年,全球痴呆患者人数已达4680万人,其中50%-75%为阿尔兹海默症患者。目前在中国,阿尔兹海默症患者人数已居世界第一,同时也是全球增速最快的国家之一。然而,阿尔兹海默症的诊疗

【相似文献】