摘要
对数字化书法作品检索的相关理论、方法及技术进行研究,提出一种新的融合余弦度量的形状匹配算法,通过把轮廓点间的余弦度量融合到欧氏距离中,更准确地反映出图像数据点之间的空间分布情况。试验结果表明,相对于采用基于欧氏距离的单一相似性度量匹配方法,融合余弦度量的形状匹配方法更能提高书法字检索的准确率。
This paper proposes a new similarity measurement method based on the combination of Euclidean distance and angle cosine to better decide the spatial distribution of contour points set in a high dimensional space so as to improve the retrieval of calligraphy works.The experiments show that the proposed method is capable of recognize calligraphy characters with higher reliability than that the conventional similarity measurement solely based on Euclidean distance can achieve.
作者
张龙海
章夏芬
韩德志
毕坤
ZHANG Longhai ZHANG Xiafen HAN Dezhi BI Kun(China Shipping Network Technology Co.,Ltd.,Information Engineering College,Shanghai Maritime University,Shanghai 201306, China)
出处
《上海船舶运输科学研究所学报》
2016年第3期84-88,共5页
Journal of Shanghai Ship and Shipping Research Institute
基金
国家自然科学基金(61303100)
上海海事大学校基金(20130467)
关键词
欧氏距离
夹角余弦
书法字
数字图书馆
Euclidean distance
included angle cosine
Chinese calligraphy
digital library
