High-Speed Tracking with Kernelized Correlation Filters

High-Speed Tracking with Kernelized Correlation Filters

The core component of most modern trackers is a discriminative classifier, tasked with distinguishing between the target and the surrounding environment. To cope with natural image changes, this classifier is typically trained with translated and scaled sample patches. Such sets of samples are riddl...

Ausführliche Beschreibung

Bibliographische Detailangaben
Veröffentlicht in:IEEE transactions on pattern analysis and machine intelligence. - 1979. - 37(2015), 3 vom: 01. März, Seite 583-96
1. Verfasser: Henriques, João F (VerfasserIn)
Weitere Verfasser: Caseiro, Rui, Martins, Pedro, Batista, Jorge
Format: Online-Aufsatz
Sprache:English
Veröffentlicht: 2015
Zugriff auf das übergeordnete Werk:IEEE transactions on pattern analysis and machine intelligence
Schlagworte:Journal Article Research Support, Non-U.S. Gov't
LEADER 01000caa a22002652 4500
001 NLM252591399
003 DE-627
005 20250219030423.0
007 cr uuu---uuuuu
008 231224s2015 xx |||||o 00| ||eng c
024 7 |a 10.1109/TPAMI.2014.2345390  |2 doi 
028 5 2 |a pubmed25n0841.xml 
035 |a (DE-627)NLM252591399 
035 |a (NLM)26353263 
040 |a DE-627  |b ger  |c DE-627  |e rakwb 
041 |a eng 
100 1 |a Henriques, João F  |e verfasserin  |4 aut 
245 1 0 |a High-Speed Tracking with Kernelized Correlation Filters 
264 1 |c 2015 
336 |a Text  |b txt  |2 rdacontent 
337 |a ƒaComputermedien  |b c  |2 rdamedia 
338 |a ƒa Online-Ressource  |b cr  |2 rdacarrier 
500 |a Date Completed 24.11.2015 
500 |a Date Revised 10.09.2015 
500 |a published: Print 
500 |a Citation Status PubMed-not-MEDLINE 
520 |a The core component of most modern trackers is a discriminative classifier, tasked with distinguishing between the target and the surrounding environment. To cope with natural image changes, this classifier is typically trained with translated and scaled sample patches. Such sets of samples are riddled with redundancies-any overlapping pixels are constrained to be the same. Based on this simple observation, we propose an analytic model for datasets of thousands of translated patches. By showing that the resulting data matrix is circulant, we can diagonalize it with the discrete Fourier transform, reducing both storage and computation by several orders of magnitude. Interestingly, for linear regression our formulation is equivalent to a correlation filter, used by some of the fastest competitive trackers. For kernel regression, however, we derive a new kernelized correlation filter (KCF), that unlike other kernel algorithms has the exact same complexity as its linear counterpart. Building on it, we also propose a fast multi-channel extension of linear correlation filters, via a linear kernel, which we call dual correlation filter (DCF). Both KCF and DCF outperform top-ranking trackers such as Struck or TLD on a 50 videos benchmark, despite running at hundreds of frames-per-second, and being implemented in a few lines of code (Algorithm 1). To encourage further developments, our tracking framework was made open-source 
650 4 |a Journal Article 
650 4 |a Research Support, Non-U.S. Gov't 
700 1 |a Caseiro, Rui  |e verfasserin  |4 aut 
700 1 |a Martins, Pedro  |e verfasserin  |4 aut 
700 1 |a Batista, Jorge  |e verfasserin  |4 aut 
773 0 8 |i Enthalten in  |t IEEE transactions on pattern analysis and machine intelligence  |d 1979  |g 37(2015), 3 vom: 01. März, Seite 583-96  |w (DE-627)NLM098212257  |x 1939-3539  |7 nnns 
773 1 8 |g volume:37  |g year:2015  |g number:3  |g day:01  |g month:03  |g pages:583-96 
856 4 0 |u http://dx.doi.org/10.1109/TPAMI.2014.2345390  |3 Volltext 
912 |a GBV_USEFLAG_A 
912 |a SYSFLAG_A 
912 |a GBV_NLM 
912 |a GBV_ILN_350 
951 |a AR 
952 |d 37  |j 2015  |e 3  |b 01  |c 03  |h 583-96