Presentation 2006-03-16
Directed NFTG incorporating Chamfer Matching and Its application
Daiki Okada, Toshikazu Wada, Junya Sakagaito,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) We have proposed an image retrieval based object tracking and recognition method. This method consists of two techniques: Pose Parameter Embedding (PPE) and approximate nearest neighbor search based on Nearest First Traversing Graph (NFTG). PPE is a technique for image retrieval based object tracking, which stores misaligned object image with pose correction parameters for updating object position, scale, and orientation. PPE requires many misaligned images stored in the image database, and it reduces the scalability of the method. In this paper, we propose an extended method incorporating Chamfer matching with the NFTG based NN search. Since our method does not use PPE, it drastically reduces the number of images in the database. Also, it improves the stability against the illumination change, because Chamfer matching uses edge images. The asymmetric property of the Chamfer metric makes NFTG directed graph. Through the experiments, we confirmed the effectiveness of our method.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Chamfer Matching / Nearest First Traversing / NFTG / Object Tracking and Recognition
Paper # PRMU2005-246
Date of Issue

Conference Information
Committee PRMU
Conference Date 2006/3/9(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair
Vice Chair
Secretary
Assistant

Paper Information
Registration To Pattern Recognition and Media Understanding (PRMU)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Directed NFTG incorporating Chamfer Matching and Its application
Sub Title (in English)
Keyword(1) Chamfer Matching
Keyword(2) Nearest First Traversing
Keyword(3) NFTG
Keyword(4) Object Tracking and Recognition
1st Author's Name Daiki Okada
1st Author's Affiliation Department of Computer and Communication Science, Wakayama University()
2nd Author's Name Toshikazu Wada
2nd Author's Affiliation Department of Computer and Communication Science, Wakayama University
3rd Author's Name Junya Sakagaito
3rd Author's Affiliation Department of Computer and Communication Science, Wakayama University
Date 2006-03-16
Paper # PRMU2005-246
Volume (vol) vol.105
Number (no) 673
Page pp.pp.-
#Pages 8
Date of Issue