摘要
为了在二维形状上寻求具有不变特征的形状描述,从三维的角度考察二维物体的形状信息,提出基于热核的形状分类算法.首先对原始物体的边界进行采样,并将内部区域三角化;然后通过优化的方法,把二维形状转化为表面光滑且封闭的三维网格曲面;最后提取三维模型表面的热核特征,利用词袋模型得到物体的特征向量,最终实现物体的形状分类.在MPEG-7与Animal Shapes数据库上的实验结果表明,与传统算法相比,该算法分类的准确率更高,鲁棒性更强.
To seek for an isometry-invariant 2D shape descriptor, we encode 2D shapes from 3D perspective,and propose a novel shape classification approach based on heat kernel. First, we build triangulation for the regionenclosed by the contour. Then, we transform the 2D shape into a 3D closed/smooth surface through a setof optimization techniques. Finally, the heat kernel signature of the 3D counterpart is extracted to identify theoriginal 2D shape. Extensive experimental results on the MPEG-7 and Animal Shapes benchmarks exhibit anadvantage of classification in terms of accuracy and robustness.
作者
孙德超
陈双敏
周亚训
陈能仑
辛士庆
王仁芳
Sun Dechao;Chen Shuangmin;Zhou Yaxun;Chen Nenglun;Xin Shiqing;Wang Renfang(Faculty of Electrical Engineering and Computer Science,Ningbo University,Ningbo 315211;College of Electronics and Computer,Zhejiang Wanli University,Ningbo 315100;Department of Computer Science,The University of Hong Kong,Hong Kong 999077;School of Computer and Science,Shandong University,Ji’nan 250101)
出处
《计算机辅助设计与图形学学报》
EI
CSCD
北大核心
2018年第8期1431-1437,共7页
Journal of Computer-Aided Design & Computer Graphics
基金
国家自然科学基金(61772016)
浙江省科技计划项目(LGG18F020001
2016C31084)
浙江省自然科学基金(LY17F020001)
宁波大学教研项目(JYXMxsj201405)
关键词
形状分类
热核特征
三维建模
数值优化
shape classification
heat kernel feature
3D modeling
numerical optimization
