Abstract
In this paper, we present a simple yet fast and robust algorithm which exploits the dense spatio-temporal context for visual tracking. Our approach formulates the spatio-temporal relationships between the object of interest and its locally dense contexts in a Bayesian framework, which models the statistical correlation between the simple low-level features (i.e., image intensity and position) from the target and its surrounding regions. The tracking problem is then posed by computing a confidence map which takes into account the prior information of the target location and thereby alleviates target location ambiguity effectively. We further propose a novel explicit scale adaptation scheme, which is able to deal with target scale variations efficiently and effectively. The Fast Fourier Transform (FFT) is adopted for fast learning and detection in this work, which only needs 4 FFT operations. Implemented in MATLAB without code optimization, the proposed tracker runs at 350 frames per second on an i7 machine. Extensive experimental results show that the proposed algorithm performs favorably against state-of-the-art methods in terms of efficiency, accuracy and robustness.
| Original language | English |
|---|---|
| Pages (from-to) | 127-141 |
| Number of pages | 15 |
| Journal | Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) |
| Volume | 8693 LNCS |
| Issue number | PART 5 |
| DOIs | |
| Publication status | Published - 2014 Jan 1 |
| Event | 13th European Conference on Computer Vision, ECCV 2014 - Zurich, Switzerland Duration: 2014 Sep 6 → 2014 Sep 12 |
Fingerprint
All Science Journal Classification (ASJC) codes
- Theoretical Computer Science
- Computer Science(all)
Cite this
}
Fast visual tracking via dense spatio-temporal context learning. / Zhang, Kaihua; Zhang, Lei; Liu, Qingshan; Zhang, David; Yang, Ming Hsuan.
In: Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), Vol. 8693 LNCS, No. PART 5, 01.01.2014, p. 127-141.Research output: Contribution to journal › Conference article
TY - JOUR
T1 - Fast visual tracking via dense spatio-temporal context learning
AU - Zhang, Kaihua
AU - Zhang, Lei
AU - Liu, Qingshan
AU - Zhang, David
AU - Yang, Ming Hsuan
PY - 2014/1/1
Y1 - 2014/1/1
N2 - In this paper, we present a simple yet fast and robust algorithm which exploits the dense spatio-temporal context for visual tracking. Our approach formulates the spatio-temporal relationships between the object of interest and its locally dense contexts in a Bayesian framework, which models the statistical correlation between the simple low-level features (i.e., image intensity and position) from the target and its surrounding regions. The tracking problem is then posed by computing a confidence map which takes into account the prior information of the target location and thereby alleviates target location ambiguity effectively. We further propose a novel explicit scale adaptation scheme, which is able to deal with target scale variations efficiently and effectively. The Fast Fourier Transform (FFT) is adopted for fast learning and detection in this work, which only needs 4 FFT operations. Implemented in MATLAB without code optimization, the proposed tracker runs at 350 frames per second on an i7 machine. Extensive experimental results show that the proposed algorithm performs favorably against state-of-the-art methods in terms of efficiency, accuracy and robustness.
AB - In this paper, we present a simple yet fast and robust algorithm which exploits the dense spatio-temporal context for visual tracking. Our approach formulates the spatio-temporal relationships between the object of interest and its locally dense contexts in a Bayesian framework, which models the statistical correlation between the simple low-level features (i.e., image intensity and position) from the target and its surrounding regions. The tracking problem is then posed by computing a confidence map which takes into account the prior information of the target location and thereby alleviates target location ambiguity effectively. We further propose a novel explicit scale adaptation scheme, which is able to deal with target scale variations efficiently and effectively. The Fast Fourier Transform (FFT) is adopted for fast learning and detection in this work, which only needs 4 FFT operations. Implemented in MATLAB without code optimization, the proposed tracker runs at 350 frames per second on an i7 machine. Extensive experimental results show that the proposed algorithm performs favorably against state-of-the-art methods in terms of efficiency, accuracy and robustness.
UR - http://www.scopus.com/inward/record.url?scp=84906482212&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84906482212&partnerID=8YFLogxK
U2 - 10.1007/978-3-319-10602-1_9
DO - 10.1007/978-3-319-10602-1_9
M3 - Conference article
AN - SCOPUS:84906482212
VL - 8693 LNCS
SP - 127
EP - 141
JO - Lecture Notes in Computer Science
JF - Lecture Notes in Computer Science
SN - 0302-9743
IS - PART 5
ER -