%0 Conference Proceedings
%4 sid.inpe.br/sibgrapi/2016/08.01.01.59
%2 sid.inpe.br/sibgrapi/2016/08.01.01.59.50
%@doi 10.1109/SIBGRAPI.2016.024
%T Efficient object recognition using sampling of keypoint triples and keygraph structure
%D 2016
%A Dazzi, Estephan,
%A De Campos, Teófilo,
%A Hilton, Adrian,
%A Cesar Jr., Roberto Marcondes,
%@affiliation Instituto de Matemática e Estatística - Universidade de São Paulo
%@affiliation CVSSP - University of Surrey
%@affiliation CVSSP - University of Surrey
%@affiliation Instituto de Matemática e Estatística - Universidade de São Paulo
%E Aliaga, Daniel G.,
%E Davis, Larry S.,
%E Farias, Ricardo C.,
%E Fernandes, Leandro A. F.,
%E Gibson, Stuart J.,
%E Giraldi, Gilson A.,
%E Gois, João Paulo,
%E Maciel, Anderson,
%E Menotti, David,
%E Miranda, Paulo A. V.,
%E Musse, Soraia,
%E Namikawa, Laercio,
%E Pamplona, Mauricio,
%E Papa, João Paulo,
%E Santos, Jefersson dos,
%E Schwartz, William Robson,
%E Thomaz, Carlos E.,
%B Conference on Graphics, Patterns and Images, 29 (SIBGRAPI)
%C São José dos Campos, SP, Brazil
%8 4-7 Oct. 2016
%I IEEE Computer Society´s Conference Publishing Services
%J Los Alamitos
%S Proceedings
%K Local image feature matching, semi-local graph matching, graph topological properties.
%X We present an object matching method that employs matches of local graphs of keypoints, called keygraphs, instead of simple keypoint matches. For a keygraph match to be valid, vertex (keypoint) descriptors must be similar and both keygraphs must satisfy structural properties concerning keypoints orientation, scale, relative position and cyclic ordering; as a result, the large majority of initial incorrect keypoint matches is correctly filtered out. We introduce a novel approach to sample keypoint triples (i.e. keygraphs) in a query image, based on complementary Delaunay triangulations; this generates a linear number of triples with relation to the number of keypoints. Query keygraphs are then matched against the indexed model keypoints; each established keygraph match is used to evaluate a candidate pose (an affine transformation). The proposed method has been evaluated for object recognition and pose estimation, achieving a better performance in comparison to state-of-the-art methods.
%@language en
%3 84.pdf