TY - GEN
T1 - Symbiotic black-box tracker
AU - Zhang, Longfei
AU - Gao, Yue
AU - Hauptmann, Alexander
AU - Ji, Rongrong
AU - Ding, Gangyi
AU - Super, Boaz
PY - 2012
Y1 - 2012
N2 - Many trackers have been proposed for tracking objects individually in previous research. However, it is still difficult to trust any single tracker over a variety of circumstances. Therefore, it is important to estimate how well each tracker performs and fusion the tracking results. In this paper, we propose a symbiotic black-box tracker (SBB) that learns only from the output of individual trackers, which run in parallel, without any detailed information about these trackers and selects the best one to generate the tracking result. All trackers are considered as black-boxes and SBB learns the best combination scheme for all existing tracking results. SBB estimates confidence scores of these trackers. The confidence score is estimated based on the tracking performance of each tracker and the consistency performance among different trackers. SBB is employed to select the best tracker with the maximum confidence score. Experiments and comparisons conducted on the "Caremedia" dataset and the "Caviar" dataset demonstrate the effectiveness of the proposed method.
AB - Many trackers have been proposed for tracking objects individually in previous research. However, it is still difficult to trust any single tracker over a variety of circumstances. Therefore, it is important to estimate how well each tracker performs and fusion the tracking results. In this paper, we propose a symbiotic black-box tracker (SBB) that learns only from the output of individual trackers, which run in parallel, without any detailed information about these trackers and selects the best one to generate the tracking result. All trackers are considered as black-boxes and SBB learns the best combination scheme for all existing tracking results. SBB estimates confidence scores of these trackers. The confidence score is estimated based on the tracking performance of each tracker and the consistency performance among different trackers. SBB is employed to select the best tracker with the maximum confidence score. Experiments and comparisons conducted on the "Caremedia" dataset and the "Caviar" dataset demonstrate the effectiveness of the proposed method.
KW - Object tracking
KW - data association
KW - information propagation
KW - multi-tracker
UR - http://www.scopus.com/inward/record.url?scp=84862924531&partnerID=8YFLogxK
U2 - 10.1007/978-3-642-27355-1_14
DO - 10.1007/978-3-642-27355-1_14
M3 - Conference contribution
AN - SCOPUS:84862924531
SN - 9783642273544
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 126
EP - 137
BT - Advances in Multimedia Modeling - 18th International Conference, MMM 2012, Proceedings
T2 - 18th International Conference on Multimedia Modeling, MMM 2012
Y2 - 4 January 2012 through 6 January 2012
ER -