Early action proposal consists in generating high quality candidate temporal segments that are likely to contain an action in a video stream, as soon as they happen. Many sophisticated approaches have been proposed for the action proposal problem but from the off-line perspective. On the contrary, we focus on the on-line version of the problem, proposing a simple classifier-based model, using standard 3D CNNs, that performs significantly better than the state of the art.
Cite
@article{arxiv.1810.07420,
title = {Embarrassingly Simple Model for Early Action Proposal},
author = {Marcos Baptista-Ríos and Roberto J. López-Sastre and Franciso Javier Acevedo-Rodríguez and Saturnino Maldonado-Bascón},
journal= {arXiv preprint arXiv:1810.07420},
year = {2018}
}
Comments
Published in the Anticipating Human Behavior Workshop, ECCV 2018