一种基于图形识别的甲骨文分类方法  

A graphic-based method for Chinese Oracle-bone classification

作　　者：吕肖庆[1,2] 李沫楠[1] 蔡凯伟[1] 王晓[1] 唐英敏[1,2] 

机构地区：[1]北京大学计算机科学技术研究所,北京100871 [2]中国文字字体设计与研究中心,北京100871

出　　处：《北京信息科技大学学报（自然科学版）》2010年第S2期92-96,共5页Journal of Beijing Information Science and Technology University

摘　　要：作为中国早期的象形文字,甲骨文的识别一直是一个难题。为探索与之相关的甲骨文自动分类问题,提出了一种基于曲率直方图的傅里叶描述子——FDCH,并将这种特征进一步用于对甲骨文文字的分类。该方法具有平移、旋转、尺度不变性。小范围的实验表明,这种方法能够较好地实现对甲骨文文字的自动分类。The Chinese oracle-bone inscription is one of the earliest known hieroglyphic inscriptions,and it has been a difficult problem for decades to automatically recognize the inscription.In order to solve the relative classification tasks,this paper proposed a Fourier Descriptor based on Curvature Histogram(FDCH),and further employed it to classify oracle-bone inscription,which can be adopted to improve the classification as the retrieval result of the oracle-bone inscription.The proposed descriptor is of translation,rotation and scaling invariance.Experiment within small scope indicates that this method performs well in inscription classification.

关 键 词：曲率直方图 傅里叶描述子 甲骨文识别 甲骨文分类 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]