TY - GEN
T1 - Action Recognition for Videos by Long-Term Point Trajectory Analysis with Background Removal
AU - Xiang, Yu
AU - Okada, Yoshihiro
AU - Kaneko, Kosuke
N1 - Publisher Copyright:
© 2016 IEEE.
PY - 2017/4/21
Y1 - 2017/4/21
N2 - Recently, dense trajectories were shown to be an efficient video motion representation for action recognition and achieved state-of-the-art results on a variety of video datasets. This paper improves their performance by taking into account camera motion. To estimate camera motion, the authors use long-term point trajectory analysis to cluster image points and propose an algorithm to find possible background cluster from these clusters according to background nature in a video. Considering the original clusters could not segment the foreground and background very well. The authors optimize the background cluster, and use the cluster to rectify the trajectory. Experimental results on three challenging action datasets (i.e., Hollywood2, Olympic Sports and UCF50) show that the rectified trajectories significantly outperform original dense trajectories.
AB - Recently, dense trajectories were shown to be an efficient video motion representation for action recognition and achieved state-of-the-art results on a variety of video datasets. This paper improves their performance by taking into account camera motion. To estimate camera motion, the authors use long-term point trajectory analysis to cluster image points and propose an algorithm to find possible background cluster from these clusters according to background nature in a video. Considering the original clusters could not segment the foreground and background very well. The authors optimize the background cluster, and use the cluster to rectify the trajectory. Experimental results on three challenging action datasets (i.e., Hollywood2, Olympic Sports and UCF50) show that the rectified trajectories significantly outperform original dense trajectories.
UR - http://www.scopus.com/inward/record.url?scp=85019242572&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85019242572&partnerID=8YFLogxK
U2 - 10.1109/SITIS.2016.13
DO - 10.1109/SITIS.2016.13
M3 - Conference contribution
AN - SCOPUS:85019242572
T3 - Proceedings - 12th International Conference on Signal Image Technology and Internet-Based Systems, SITIS 2016
SP - 23
EP - 30
BT - Proceedings - 12th International Conference on Signal Image Technology and Internet-Based Systems, SITIS 2016
A2 - De Pietro, Giuseppe
A2 - Dipanda, Albert
A2 - Chbeir, Richard
A2 - Gallo, Luigi
A2 - Yetongnon, Kokou
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 12th International Conference on Signal Image Technology and Internet-Based Systems, SITIS 2016
Y2 - 28 November 2016 through 1 December 2016
ER -