文章摘要

文章信息/Info

Author(s):: CHEN Jing; JIN Chen; TENG Shenghua; College of Electronics, Communication and Physics, Shandong University of Science and Technology, Qingdao 266590, China

Keywords:: Alzheimer′s disease; Magnetic resonance imaging; Gaussian process; Stepwise classification; Kullback-leibler divergence

摘要:: 脑影像数据维数高且有效训练样本少是影响阿尔茨海默病计算机辅助诊断性能的重要因素。对此小样本分类问题，以高斯过程为基础设计了一种分步的分类方法：先对测试样本利用高斯过程进行初步分类；依据后验概率筛选类别归属确定性强的样本作为补充参与训练，再对其余错分可能性相对较高的样本重新进行分类。利用ADNI数据库磁共振影像的分类实验表明，二次分类倾向于增大样本归属于真实类别的后验概率、提高类别判定的确定性，分类性能优于常规的高斯过程分类方法和支持向量机。

Abstract:: The computer-aided diagnosis of Alzheimer’s disease from brain imaging generally is affected by high data dimensionality and the lack of training samples. On the basis of the Gaussian process, a stepwise classification frame work was designed to alleviate this small-sample problem. All test samples were firstly classified by Gaussian process. The samples with high or low posterior probabilities were identified as being correctly classified with high confidence, and then included into the training data to reclassify the rest samples. Experiments on ADNI database show that the second classification tends to increase the posterior probability of the test sample belonging to the right category and improve the classification certainty. The classification performance of the proposed method is superior to the conventional Gaussian process and support vector machine (SVM).