Unified Real-Time Tracking and Recognition with Rotation-Invariant Fast Features

被引：63

作者：

Takacs, Gabriel ^{[1
]}

Chandrasekhar, Vijay ^{[1
]}

Tsai, Sam ^{[1
]}

Chen, David ^{[1
]}

Grzeszczuk, Radek ^{[2
]}

Girod, Bernd ^{[1
]}

机构：

[1] Stanford Univ, Informat Syst Lab, Stanford, CA 94305 USA

[2] Nokia Res Ctr, Palo Alto, CA USA

来源：

2010 IEEE CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR) | 2010年

关键词：

D O I：

10.1109/CVPR.2010.5540116

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

We present a method that unifies tracking and video content recognition with applications to Mobile Augmented Reality (MAR). We introduce the Radial Gradient Transform (RGT) and an approximate RGT, yielding the Rotation-Invariant, Fast Feature (RIFF) descriptor. We demonstrate that RIFF is fast enough for real-time tracking, while robust enough for large scale retrieval tasks. At 26x the speed, our tracking-scheme obtains a more accurate global affine motion-model than the Kanade Lucas Tomasi (KLT) tracker. The same descriptors can achieve 94% retrieval accuracy from a database of 10(4) images.

引用

页码：934 / 941

页数：8

共 28 条

[1] Ahonen T, 2009, LECT NOTES COMPUTER, V5575
[2] [Anonymous], 2006, PROC EUR C COMPUT VI
[3] Banerjee A, 2004, SIAM PROC S, P234
[4] Birchfield S, 2007, Klt: An implementation of the Kanade-Lucas-Tomasi feature tracker
[5] BRASNETT P, 2007, VIE JUL
[6] Calonder M., 2009, INT C COMP VIS ICCV
[7] Chandrasekhar V., 2009, CVPR
[8] CHANDRASEKHAR V, 2009, CVPR MIAM FLOR JUN
[9] Chen D. M., 2008, CD COVER DATABASE QU
[10] Jegou H, 2008, LECT NOTES COMPUT SC, V5302, P304, DOI 10.1007/978-3-540-88682-2_24

← 1 2 3 →