指纹分类是针对大型指纹库的一个重要的索引方式, 可以有效地提高指纹匹配的效率. 指纹类型的不同表现为指纹纹

理结构的差异, 而指纹的方向场则可以有效地描述纹理结构的差异. 同一类型指纹不同区域上方向角结构的差异以及相邻区域

间方向角结构的联系可以视作一个马尔可夫随机场. 本文利用嵌入式隐马尔可夫模型对指纹方向场进行建模分析, 通过合理地

抽取指纹的类型特征, 构造观察向量、进行建模训练, 然后利用训练好的马尔可夫模型进行匹配, 最终提出并实现了一种新的鲁

棒性强且精度较高的指纹分类方法.-Fingerprint classification is an important index for a large-scale fingerprint database, can effectively improve the efficiency of the fingerprint matching, the different performance of the fingerprint type fingerprint texture differences in the structure, while the direction of the field of the fingerprint can be effectively describe the texture differences in the structure. fingerprint different regions of the same type the direction angle differences in the structure and the adjacent areas between the direction of the angular structure of the contact can be regarded as a Markov Random Field paper embedded hidden Markov model fingerprint orientation field modeling analysis by reasonable to extract fingerprint type characteristics, construct observation vectors, modeling training, and then take advantage of the training good match Markov model eventually proposed and implemented a new robust and high precision fingerprint classification .