Weakly Supervised Action Learning with RNN based Fine-to-coarse Modeling

Alexander Richard, Hilde Kuehne, Juergen Gall

URL

Proceedings: CVPR

Year: 2017

Peer Reviewed : Yes

Results: 
Breakfast Dataset /BF Action seg. weakly (MoF): 33.3% (GRU + reestimation)
Breakfast Dataset /BF Action alignment weakly (IoD): 47.3% (GRU + reestimation)
Hollywood Extended /HE Action alignment weakly (IoD): 51.1% (GRU + reestimation )