基于结构图的手绘草图检索
高竹红,
(安徽大学 摘
汤
进,
罗
斌
计算智能与信号处理教育部重点实验室,安徽 合肥 230039)
要:提出一种基于结构图的草图检索方法,在保持图像大致形状的基础上对图像的轮廓数据进行有效
的降维处理,解决了轮廓的起点选择,旋转,平移以及缩放等不变性,形成了图像的标准化形状描述子, 并生成相应的结构图,提取结构图的四种特征,分别进行形似性度量.实验表明文中方法能够用来有效地 进行草图检索. 关键词:结构图;草图检索;标准化形状描述子;图谱特征 中图分类号:TP391.41 文献标识码:A 文章编号:1673-629X(2008)03-0032-04
Sketch Retrieval Based on Structural Graph
GAO Zhu-hong, TANG Jin, LUO Bin Hefei (Ministry of Education Key Lab. of Intelligent Computing and Signal Processing , Anhui University, 230039, China) Abstract: A sketch retrieval method is proposed based on structural graph representation in this paper. The dimensionality of contour representation is reduced while the shape of the sketch is reasonably maintained. It is invariant while starting point is different or rotation, translation and scaling of the contours are existed. A normalized shape descriptor is constructed and corresponding structural graph is generated. With the graph representation in hand, four kinds of graph spectral features are calculated to compare the similarity of the graphs. Experimental results show that the proposed method can be used to retrieve sketch database effectively. Key words: structural graph; sketch retrieval; normalized shape descriptor; graph spectral features