TY - GEN
T1 - A Context-Aware Loss Function for Action Spotting in Soccer Videos
AU - Cioppa, Anthony
AU - Deliege, Adrien
AU - Giancola, Silvio
AU - Ghanem, Bernard
AU - Droogenbroeck, Marc Van
AU - Gade, Rikke
AU - Moeslund, Thomas B.
PY - 2020/6
Y1 - 2020/6
N2 - In video understanding, action spotting consists in temporally localizing human-induced events annotated with single timestamps. In this paper, we propose a novel loss function that specifically considers the temporal context naturally present around each action, rather than focusing on the single annotated frame to spot. We benchmark our loss on a large dataset of soccer videos, SoccerNet, and achieve an improvement of 12.8% over the baseline. We show the generalization capability of our loss for generic activity proposals and detection on ActivityNet, by spotting the beginning and the end of each activity. Furthermore, we provide an extended ablation study and display challenging cases for action spotting in soccer videos. Finally, we qualitatively illustrate how our loss induces a precise temporal understanding of actions and show how such semantic knowledge can be used for automatic highlights generation.
AB - In video understanding, action spotting consists in temporally localizing human-induced events annotated with single timestamps. In this paper, we propose a novel loss function that specifically considers the temporal context naturally present around each action, rather than focusing on the single annotated frame to spot. We benchmark our loss on a large dataset of soccer videos, SoccerNet, and achieve an improvement of 12.8% over the baseline. We show the generalization capability of our loss for generic activity proposals and detection on ActivityNet, by spotting the beginning and the end of each activity. Furthermore, we provide an extended ablation study and display challenging cases for action spotting in soccer videos. Finally, we qualitatively illustrate how our loss induces a precise temporal understanding of actions and show how such semantic knowledge can be used for automatic highlights generation.
UR - http://www.scopus.com/inward/record.url?scp=85094809735&partnerID=8YFLogxK
U2 - 10.1109/CVPR42600.2020.01314
DO - 10.1109/CVPR42600.2020.01314
M3 - Article in proceeding
SN - 978-1-7281-7169-2
T3 - I E E E Conference on Computer Vision and Pattern Recognition. Proceedings
SP - 13123
EP - 13133
BT - 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
PB - IEEE
T2 - 2020 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)
Y2 - 14 June 2020 through 19 June 2020
ER -