DocumentCode :
1742857
Title :
Edge orientation-based multi-view object recognition
Author :
Zhu, Weiyu ; Levinson, Stephen
Author_Institution :
Dept. of Electr. & Comput. Eng., Illinois Univ., Urbana, IL, USA
Volume :
1
fYear :
2000
fDate :
2000
Firstpage :
936
Abstract :
An edge orientation-based algorithm for multi-view object recognition is presented. The distribution of edge point orientations, combined with the normalized second moments, is taken as a feature vector to describe and index each object instance. For each unknown test object, a set of likelihood weights for all the possible candidate objects is obtained by computing the Euclidean distances between the unknown feature set and all the available template feature vectors. A convincing coefficient is introduced to evaluate the confidence of the best match. New views (photo shots) will be automatically taken if the best match is thought to be insufficiently convincing. In experiments, our algorithm has achieved an average of 91.5% correct recognition rate under the 5-view scheme for 320 testing images taken from eight natural objects
Keywords :
feature extraction; object recognition; visual databases; Euclidean distances; edge orientation-based algorithm; edge point orientations; feature vector; likelihood weights; multi-view object recognition; natural objects; normalized second moments; recognition rate; template feature vectors; unknown feature set; Computer vision; Decision trees; Educational robots; Feature extraction; Image recognition; Lighting; Object oriented databases; Object recognition; Spatial databases; Testing;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Pattern Recognition, 2000. Proceedings. 15th International Conference on
Conference_Location :
Barcelona
ISSN :
1051-4651
Print_ISBN :
0-7695-0750-6
Type :
conf
DOI :
10.1109/ICPR.2000.905600
Filename :
905600
Link To Document :
بازگشت