基于HMM和SVM的指纹分类方法

FINGERPRINT CLASSIFICATION USING HMM AND SVM

该文提出了指纹分类的一种新型方法:使用指纹编码的基于隐马尔可夫模型(HMM)和支持向量机(SVM)的两级分类,该方法采用FingerCode作为指纹的特征表述,首先用5个伪二维 HMM进行类别初选,确定最可能的两种指纹分类结果,再用相应的 SVM分类器做最终判决,实验表明,分类性能已经达到或超过目前流行的指纹分类算法,具有一定的实用价值。

In this paper, a new classification method is proposed that is based on HMM and SVM using FingerCode. This technique uses FingerCode as representation of fingerprint, first, using five HMM classifiers to classify originally, furthermore, basing on the most possible two results of classification by HMM, the corresponding SVM classifier is selected to make final decision. Through experiments, the performance of this method reaches or exceeds many popular ones and has value to a certainty in practice.