Research on Video Emotion Recognition Based on Attention Mechanism LSTM Model

DOI

10.2991/ncce-18.2018.149How to use a DOI?

Keywords

human-computer; LSTM; CHEAVD; CNN; videos.

Abstract

The study of video-based emotion recognition is still a difficult and hot research issue in the area of computer vision and human-computer interaction. How to effectively extract key frames from videos, construct spatiotemporal feature space of videos, and build model based on the mapping relationship between temporal feature space and video emotion type space, has become an important issue of video-based emotion recognition study. To solve this problem, this paper proposes a video emotion recognition method based on the attention mechanism LSTM model. Based on the features extracted by CNN, this method uses the LSTM model based on attention mechanism to model video temporal feature space and construct video emotion recognition model. The experimental results on CHEAVD data set show that this method can effectively improve the recognition rate of video emotion recognition task.

Copyright

© 2018, the Authors. Published by Atlantis Press.

Open Access

This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)

TY  - CONF
AU  - Xiaobin Zheng
PY  - 2018/05
DA  - 2018/05
TI  - Research on Video Emotion Recognition Based on Attention Mechanism LSTM Model
BT  - Proceedings of the 2018 International Conference on Network, Communication, Computer Engineering (NCCE 2018)
PB  - Atlantis Press
SP  - 894
EP  - 898
SN  - 1951-6851
UR  - https://doi.org/10.2991/ncce-18.2018.149
DO  - 10.2991/ncce-18.2018.149
ID  - Zheng2018/05
ER  -