|本期目录/Table of Contents|

[1]杨小军,杨兴炜,曾  峦,等. 基于轮廓关键点集的形状分类 [J].南京大学学报(自然科学),2010,46(1):47-55.
 Yang Xiao J un,Yang Xing Wei,Zeng L uan,et al. Shape classification using contour critical point sets [J].Journal of Nanjing University(Natural Sciences),2010,46(1):47-55.
点击复制

 基于轮廓关键点集的形状分类

()
     

《南京大学学报(自然科学)》[ISSN:0469-5097/CN:32-1169/N]

卷:
46
期数:
2010年第1期
页码:
47-55
栏目:
出版日期:
2010-01-30

文章信息/Info

Title:
 Shape classification using contour critical point sets

作者:
 杨小军 1 杨兴炜 2 曾  峦 3 刘文予 4
 (1. 装备指挥技术学院研究生管理大队 , 北京 ,101416 ;2. 美国 Temple 大学计算机与信息科学系 ,
美国 , 费城 ,PA 19122 ;3. 装备指挥技术学院国防重点实验室 , 北京 ,101416 ;
4. 华中科技大学电子信息工程系 , 武汉 ,430074)
Author(s):
 Yang Xiao 2 J un 1 Yang Xing 2 Wei 2 Zeng L uan 3 L iu Wen 2 Yu 4
 (1. Company of Postgraduate Management , the Academy of Equipment Command and Technology , Beijing ,
101416 , China ;2. Department of Computer and Information Sciences , Temple University , Philadelphia ,
PA 19122 , USA ;3. Key Lab of National Defense , the Academy of Equipment Command and
Technology , Beijing , 101416 , China ;4. Department of Electronics and Information Engineering ,
Huazhong University of Science and Technology ,Wuhan , 430074 , China)
关键词:
  形状分类 轮廓关键点集 inner 2 distance 形状上下文 贝叶斯分类器
Keywords:
 shape classification contour critical point sets inner 2 distance shape context Bayesian classifier
分类号:
-
DOI:
-
文献标志码:
-
摘要:
 形状分析是计算机视觉领域的经典问题 , 目前已有大量关于形状分类问题的研究 . 但是 , 当处理大的非线性失真、 特别是结构上或者关联上的失真时 , 许多形状分类方法往往无能为力 . 提出一种
利用轮廓关键点集 (contour critical point sets ,CCPS) 进行形状分类的新方法 . 轮廓关键点的特征用其inner 2 distance 形状上下文 (IDSC) 表征 . 关键点的 inner 2 distance 形状上下文不仅表征形状的局部特征 ,
也反映其全局特征 , 这种局部点的全局特征信息对遮挡、 非线性失真等有良好的鲁棒性 . 巧妙地构造关键点的特征向量后 , 对形状轮廓关键点集、 形状类、 和全体形状样本建模 , 进行三级的贝叶斯分类 . 形状
类模型使得可以利用同一类中的不同样本的不同关键点对输入形状进行识别 . 实验结果表明 , 这种基于视觉部分的全局特征 , 三级的贝叶斯分类方法对非线性失真、 类内变异、 结构变化、 遮挡等具有良好
的鲁棒性 . 文中的方法在 Kimia 形状数据库上达到 100 % 的分类精度 , 并且分类所有 108 个测试形状仅需要 8 s , 是目前已知最好的分类性能 . 在广泛使用的 MPEG 2 7 形状数据库上 , 也能达到满意的分类结果 .
Abstract:
 Shape analysis has been one of the most studied topics in computer vision. One major task in shape analysis is to study the underlying statistics of shape population and use the information to extract , recognize , and
understand physical structures and biological objects. Matching based algorithms perform classification , essentially through exemplar based or nearest neighborhood approach by matching the query shape against all those in the
training set. On few training samples , these algorithms are hard to capture the large intra - class variation. On largetraining samples , it is extremely time consuming to perform shape matching one - by - one. Approaches based on
generative models require a large number of parameters , which renders them significantly more expensive computationally , and also increases the possibility of converging to non - optimal local minima. Furthermore , existing
Matching based and model - based approaches cannot handle object classes that have different parts or numbers of parts without splitting the class into separate subclasses. Most of the methods for shape classification are based on
contour and many researchers have worked on the general shape classification problem. However , approaches for classifying contour shapes can encounter difficulties when dealing with classes that have large nonlinear variability ,
especially when the variability is structural or due to articulation. A novel method , using contour critical point sets (CCPS) to perform shape classification task , is proposed in this paper. First , inner - distance shape context (IDSC)
is used to characterize the critical points. Of course , other features of the critical points may instead of IDSC. Shapes are represented by a set of points sampled from the shape contours and the shape context at a reference point
captures the distribution of the remaining points relative to it , thus offering a globally discriminative characterization. Corresponding points on two similar shapes will have similar shape contexts. The inner -distance is
defined as the length of the shortest path between landmark points within the shape silhouette. It is articulation insensitive and more effective at capturing part structures than the Euclidean distance. This suggests that the inner -
distance can be used as a replacement for the Euclidean distance to build more accurate descriptors for complex shapes , especially for those with articulated parts. Humans perception of shape is based on similarity of common
parts , to the extent that a single , significant visual part is sufficient to recognize the whole object and part -based representations allow for recognition that is robust in the presence of occlusion , movement , deletion , or growth of
portions of an object. It is a simple and natural observation that maximal convex or concave parts of objects determine visual parts. So the contour critical point sets (CCPS) of shapes is utilized to perform shape classification
task. The IDSC of critical point is an excellent feature of contour point , which not only contains local features but also the global information. After design the smart feature of shapes , then , Bayesian classification is performed
within a three - level framework which consists of models for contour critical point sets , for classes , and for the entire database of training examples. The class model enables different critical points of different exemplars of one class to
contribute to the recognition of an input shape. This new method achieves 100 % classification accuracy on Kimia database. Furthermore , to classify all 108 test shapes only need 8 seconds , which is the best performance ever
reported in the literature. The results on the well 2 known MPEG7 CE 2 Shape -1 data set also prove its superiority.

参考文献/References:

 [ 1 ]   Belongie S , Malik J , Puzicha J. Shape matching and object recognition using shape contexts. IEEE Transactions on Pattern Analysis and Ma -chine Intelligence , 2002 , 24(4) : 509 ~ 522.
[ 2 ]   Latecki L J , Lakamper R , Eckhardt U. Shape descriptors for non -rigid shapes with a single closed contour. Proceedings of IEEE Conference on Computer Vision and Pattern Recognition , 2000(1) : 424 ~ 429.
[ 3 ]   Sun K, Super B. Classification of contour shapes using class segment sets. Proceedings of IEEE Conference on Computer Vision and Pat -tern Recognition , 2005(2) : 727 ~ 733.
[ 4 ]   Latecki L J , Lakaemper R , Wolter D. Optimal partial shape similarity. Image and Vision Com -puting Journal , 2005 , 23 : 227 ~ 236.
[ 5 ]   McNeill G, Vijayakumar S. Part 2 based probabi -listic point matching using equivalence con -straints. The 20 th Annual Conference on Neural Information Processing Systems. Vancouver , British Columbia , Canada , 2006 , 969 ~ 976.
[ 6 ]   Tu Z, Yuille A L. Shape matching and recogni -tion using generative models and informative features. European Conference on Computer Vi -sion , 2004(3023) :195 ~ 209.
[ 7 ]   McNeill G, Vijayakumar S. A probabilistic ap -proach to robust shape matching. IEEE Interna -tional Conference on Image Processing , Atlan -ta , 2006 , 937 ~ 940.
[ 8 ]   Chui H , Rangarajan A. A new point matching algorithm for non - rigid registration. Computer Vision and Image Understanding , 2003 , 89 : 114 ~ 141.
[ 9 ]   Kaleem S , Benjamin B K. Parts of visual form: Computational aspects. IEEE Transactions on · 4 5 ·? 1994-2010 China Academic Journal Electronic Publishing House. All rights reserved. http://www.cnki.net
Pattern Analysis and Machine Intelligence , 1995 , 17(3) : 239 ~ 251.
[10]   Chen J , Stockman G. 3D free - form object rec -ognition using indexing by contour features. Computer Vision Image Understanding , 1998 , 71(3) : 334 ~ 355.
[11]   Yang D , Yin Y L , Zhu M Y, et al. A corner detection method based on strong noise adapta -tion. Journal of Nanjing Univesity (Natural Sci -ences) , 2008 , 44 (2) :211 ~ 218. ( 杨  栋 , 尹 义龙 , 朱明英等 . 一种大噪声自适应的角点检
测技术 . 南京大学学报 ( 自然科学 ) , 2008 , 44 (2) : 140 ~ 147) .
[12]   Wang ZJ , Ning X B , Yang X D. A new meth -od for multilevel fingerprint classification based on inflexionlike eigenvector. Journal of Nanjing
University (Natural Sciences) , 2007 , 43 (1) : 47 ~ 55. ( 王志坚 , 宁新宝 , 杨小冬 . 基于类拐点特征向量的多层次指纹分类新方法 . 南京大学学报 ( 自然科学 ) , 2007 , 43(1) :47 ~ 55) .
[13]   Latecki L , Lakamper R. Shape similarity meas -ure based on correspondence of visual parts. IEEE Transactions on Pattern Analysis and Ma -chine Intelligence , 2000 , 22 : 1185 ~ 1190.
[14]   Xiang B , Longin J L , Liu W Y. Skeleton prun -ing by contour partitioning with discrete curve evolution. IEEE Transactions on Pattern Anal -ysis and Machine Intelligence , 2007 , 29 (3) : 449 ~ 462.
[15]   Jia J , Cai L H. Fingerprint verification based on minutiae re - matching. Journal of Tsinghua Uni -versity ( Sciences and Technology) , 2006 , 46
(10) :1776 ~ 1779. ( 贾  珈 , 蔡莲红 . 基于局部细节特征的二次指纹匹配算法 . 清华大学学报 ( 自然科学 ) , 2006 , 46(10) :1776 ~ 1779) .
[16]   Wang W X, Yuan J , Zang J , et al . A finger -print matching algorithm of minutia based on lo -cal characteristic. Journal of Nanjing University
(Natural Sciences) , 2009 , 45(1) :18 ~ 23. ( 王伟希 , 袁  杰 , 臧  炅等 . 基于局部特征的点模式指纹匹配算法 . 南京大学学报 ( 自然科学 ) , 2009 , 45(1) : 18 ~ 23) .
[17]   Ling H B , David W J. Shape classification u -sing the inner - distance. IEEE Transactions on Pattern Analysis and Machine Intelligence , 2007 , 29(2) : 286 ~ 299.
[18]   Sebastian T B , Klein P N , Kimia B B. Recogni -tion of shapes by editing their shock graphs. IEEE Transactions on Pattern Analysis and Ma -chine Intelligence , 2004 , 26(5) : 550 ~ 571.
[19]   Yang X W , Bai X, Longin J L. Shape classifi -cation based on skeleton path similarity. The 6 th International Conference on Energy Minimiza -tion Methods in Computer Vision and Pattern Recognition , China , 2007.
[20]   Mohammad R D , Vincent T. Shape recognition and retrieval using string of symbols. The 5 thInternational Conference on Machine Learning and Applications , 2006 , 48 : 101 ~ 108.
[21]   Bai X, Longin J L. Path similarity skeleton graph matching. IEEE Transactions on Pattern Analysis and Machine Intelligence , 2008 , 30 (7) : 1282 ~ 1292.

相似文献/References:

备注/Memo

备注/Memo:
更新日期/Last Update: 2015-03-27