TY - GEN
T1 - Curvature feature based shape analysis
AU - Chen, Yufeng
AU - Li, Fengxia
AU - Huang, Tianyu
PY - 2008
Y1 - 2008
N2 - A novel method is presented to evaluate the similarity of shapes based on the curvature features and their distribution. Firstly the curvature information is used to define the curvature features, which are learned and searched by the proposed statistic method. Secondly the structural features of each pair are measured, so that the distribution of the curvature feature can be further measured. Taking both advantages of the local shape context analysis and the global feature distances optimization, our method can endure large nonrigid distortion and occlusion. The experiments, which have been implemented on the MPEG-7 shape database, show that this method is efficient and robust under certain shape distortion. Another experiment on the abnormal behavior detection shows its potential in shape detection, motion tracking, image retrieving and the related areas.
AB - A novel method is presented to evaluate the similarity of shapes based on the curvature features and their distribution. Firstly the curvature information is used to define the curvature features, which are learned and searched by the proposed statistic method. Secondly the structural features of each pair are measured, so that the distribution of the curvature feature can be further measured. Taking both advantages of the local shape context analysis and the global feature distances optimization, our method can endure large nonrigid distortion and occlusion. The experiments, which have been implemented on the MPEG-7 shape database, show that this method is efficient and robust under certain shape distortion. Another experiment on the abnormal behavior detection shows its potential in shape detection, motion tracking, image retrieving and the related areas.
UR - http://www.scopus.com/inward/record.url?scp=56549129500&partnerID=8YFLogxK
U2 - 10.1007/978-3-540-87442-3_52
DO - 10.1007/978-3-540-87442-3_52
M3 - Conference contribution
AN - SCOPUS:56549129500
SN - 3540874402
SN - 9783540874409
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 414
EP - 421
BT - Advanced Intelligent Computing Theories and Applications
T2 - 4th International Conference on Intelligent Computing, ICIC 2008
Y2 - 15 September 2008 through 18 September 2008
ER -