基于类集和类对的有监督流形学习的肺结节分类

Lung Nodule Classification Using Supervised Manifold Learning Based on All-Class and Pairwise-Class Feature Extraction

维数简约是肺结节分类识别问题中的关键步骤,现有的方法中都是将所有类别的数据作为一个整体进行降维,忽略了不同类别数据之间在特征子集上的差异性。本文提出了一种将类集和类对相结合的有监督流形特征抽取思想,并将之应用于肺结节的分类中,最终形成一个基于CT影像的肺结节分类系统。实验结果表明了方法的有效性。

Dimensionality reduction plays an important role in lung nodule classification,but in most of the existing methods,dimensionality is reduced with all classes being considered jointly,difference between feature subsets of different classes is ignored.In this paper,a supervised manifold feature extraction method based on fusion of all-class and pairwise-class is proposed,and an supervised multi-classifiers system of lung nodule classification is constructed.Experiments show a significant improvement in recognition accuracy.