Encouraging LSTMs to Anticipate Actions Very Early
Authors:
- Mohammad Sadegh AliakbarianMAMohammad Sadegh Aliakbarian
- Fatemeh Sadat SalehFSFatemeh Sadat Saleh
- Mathieu SalzmannMSMathieu Salzmann
- Basura FernandoBFBasura Fernando
- Lars PeterssonLPLars Petersson
- Lars AnderssonLALars Andersson
- 1 October 2017
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 280-289
Abstract
In contrast to the widely studied problem of recognizing an action given a complete sequence, action anticipation aims to identify the action from only partially available videos. As such, it is therefore key to the success of computer vision applications requiring to react as early as possible, such as autonomous navigation. In this paper, we propose a new action anticipation method that achieves high prediction accuracy even in the presence of a very small percentage of a video sequence. To this end, we develop a multi-stage LSTM architecture that leverages context-aware and action-aware features, and introduce a novel loss function that encourages the model to predict the correct class as early as possible. Our experiments on standard benchmark datasets evidence the benefits of our approach; We outperform the state-of-the-art action anticipation methods for early prediction by a relative increase in accuracy of 22.0% on JHMDB-21, 14.0% on UT-Interaction and 49.9% on UCF-101.Keywords
All Related Versions
This publication has 31 references indexed in Scilit:
- A Multi-stream Bi-directional Recurrent Neural Network for Fine-Grained Action Detection2016
- Predicting the Where and What of Actors and Actions through Online Action Localization2016
- Anticipating Visual Representations from Unlabeled Video2016
- Actionness Estimation Using Hybrid Fully Convolutional Networks2016
- Recurrent Neural Networks for driver activity anticipation via sensory-fusion architecture2016
- Learning Spatiotemporal Features with 3D Convolutional Networks2015
- Action recognition with trajectory-pooled deep-convolutional descriptors2015
- Action Localization with Tubelets from Motion2014
- Towards Understanding Action Recognition2013
- Action Recognition with Improved Trajectories2013