磁共振成像对大脑病征诊断仿真研究

Simulation Study of Brain Symptom Diagnosis by Magnetic Resonance Imaging

通过磁共振成像技术对大脑病症进行诊断,可以有效研究在老龄化过程中脑网络的一系列性质。对大脑病症的研究,首先对脑网络特征的诊断是前提,从磁共振图像中提取出纹理特征构成原始特征集,完成对大脑病症的诊断。传统方法先采用单一灰度特征分割脑网络图像,得到伪影因素对脑组织描述的影响,但忽略了从图像中提取出纹理特征构成原始特征集,导致诊断精度偏低。提出基于磁共振成像的脑网络特征诊断方法。先对磁共振时间序列进行符号动力学编码,得到脑网络具有显著性差异的脑区,提取出纹理特征构成原始特征集,利用独立分量构成脑网络特征子集,将训练样本与待分类样本都映射到特征子集所映射的独立空间中,利用特征子集对支持向量机分类器进行训练并对脑组织进行分类,完成对脑网络特征诊断。仿真证明,所提方法诊断精度较高,可以有效地实现对大脑病症的诊断。

This paper proposes a feature diagnosis method for brain network based on magnetic resonance ima- ging. First of all, the symbolic dynamic coding for magnetic resonance time series is carried out to obtain brain net- work with brain region of significant differences and extract texture feature to constitute original feature set, then fea- ture subset of brain network is constituted by using independent component, and then the training samples and the samples to be classified are mapped into the independent spaces mapped by feature subsets. At last, the SVM （sup- port vector machine） classifier is trained via the feature subset, and the brain tissue is classified to achieve symbolic diagnosis of brain network. The simulation results show that the method in this paper has high accuracy. It can effec- tively diagnose cerebral disease.