基于层次方法和PCA特征变换的宫颈细胞识别

对宫颈细胞多分类，可以自动识别出不同状态的细胞，进而为宫颈癌诊断提供科学依据。在用6种多分类算法实验后，选取支持向量机（SVM）作为基分类器，先用一对一策略（one- versus -one）训练6个分类器进行3分类，然后再训练1个2分类器，这种二层4分类方法提高了识别准确率。又考虑不同层特征模式的差异性，在保证识别性能同时，每层分类前先采用主成分分析（PCA）法将原始154维特征变换到低维空间，去除冗余特征，加快识别速度。实验证明，所提层次PCA法在宫颈细胞分类中相比6种传统多分类方法有更高的识别准确率，可达90%以上；识别速度也较普通层次法提升了21.31%。

Cervical cell recognition based on hierarchical method and PCA feature transformation

In order to recognize multi-class cervical cells automatically, a hierarchical method with PCA feature transformation is proposed and this cell recognition can provide the evidence for cervical cancer diagnosis. The cervical cell recognition is treated as a 4-class classification problem. There are two levels in this hierarchical method. First, one-versus-one strategy is used to train 6 SVM classifiers to do 3-class classification. Second, abnormal cells in one type of 3 categories are classified by a 2-class SVM. To optimize feature combination and reduce running time, a feature transformation method named PCA is adopted to transform the original feature vector into low-dimension feature space. The experiments show that the proposed hierarchical PCA recognition method is faster than hierarchical method at a ratio of 21.31%, can distinguish 4 cervical cell categories better than 6 other traditional methods and achieve above 90% accuracy.