Abstract: In order to improve the speed and the accuracy of fingerprint recognition, especially in the fingerprint database, the scale of which is increasing, fingerprint classification is very important. This paper presents a classification method based on the fusion of membership, and divides the fingerprint into five categories：whorl, arch, tended arch, reverse and reverse loop. This method takes full advantage of the fingerprint texture, adds up statistics by block, fuses the weighted membership, and highlights areas that can distinguish the type of fingerprints to improve the accuracy of classification. Experiments verify that the method has good classification performance.

Key words: Local Binary Pattern（LBP）, fusion membership, fingerprint classification, K-means

摘要： 为了提高指纹识别的速度和准确率，特别是在指纹数据库规模不断增大的情况下，指纹分类显得尤为重要。提出一种基于融合隶属度的分类方法，将指纹分为五类：斗型、弓型、帐弓型、左旋和右旋型。该方法充分利用指纹纹理信息，并进行分块统计，加权融合隶属度，突显出能区分指纹类别区域，提高了分类的准确率。实验证明该方法有良好的分类性能。

关键词: 局部二元模式（LBP）, 融合隶属度, 指纹分类, K均值