【摘要】 实现了一种指纹分类的新方法。这种方法将用256×256×8bit表示的指纹图象分成20×23个子区,以走取方向图象;然后进行松驰平滑处理以消除噪声,再进行奇异区检出,并根据它的方向分布模式确定奇异区的类型;最后根据指纹学的知识将指纹分成弓(Arch)、箕(Loop)、斗(Whorl)等7种类别。得到满意的分类精度。更多还原

【Abstract】 This paper presents a new method of figerprint classification. In this method, figerprint image with 256 × 256 × 8bit is divided into 20 × 23 subregions to obtain eight direction patterns. To eliminate noise components, relaxation smoothing method is used. Next, singular regions and feature points (core, delta points) are found. Finally, according to feature points count and knowledge of dactylography figerprints are classified into seven categories with satisfactory accuracy.更多还原