TY - GEN
T1 - Time-sensitive topic models for action recognition in videos
AU - Tavenard, Romain
AU - Emonet, Remi
AU - Odobez, Jean Marc
PY - 2013/1/1
Y1 - 2013/1/1
N2 - In this paper, we postulate that temporal information is important for action recognition in videos. Keeping temporal information, videos are represented as word ×time documents. We propose to use timesensitive probabilistic topic models and we extend them for the context of supervised learning. Our time-sensitive approach is compared to both PLSA and Bag-of-Words. Our approach is shown to both capture semantics from data and yield classification performance comparable to other methods, outperforming them when the amount of training data is low.
AB - In this paper, we postulate that temporal information is important for action recognition in videos. Keeping temporal information, videos are represented as word ×time documents. We propose to use timesensitive probabilistic topic models and we extend them for the context of supervised learning. Our time-sensitive approach is compared to both PLSA and Bag-of-Words. Our approach is shown to both capture semantics from data and yield classification performance comparable to other methods, outperforming them when the amount of training data is low.
UR - https://www.scopus.com/pages/publications/84897818190
U2 - 10.1109/ICIP.2013.6738615
DO - 10.1109/ICIP.2013.6738615
M3 - Conference contribution
AN - SCOPUS:84897818190
SN - 9781479923410
T3 - 2013 IEEE International Conference on Image Processing, ICIP 2013 - Proceedings
SP - 2988
EP - 2992
BT - 2013 IEEE International Conference on Image Processing, ICIP 2013 - Proceedings
PB - IEEE Computer Society
T2 - 2013 20th IEEE International Conference on Image Processing, ICIP 2013
Y2 - 15 September 2013 through 18 September 2013
ER -