Speech based deception detection using deep learning is one of the technologies to realize a deception detection system with high recognition rate in the future. Multi-network feature extraction technology can effectively improve the recognition performance of the system, but due to the limited labeled data and the lack of effective feature fusion methods, the performance of the network is limited. Based on this, a novel hybrid network model based on attentional multi-feature fusion (HN-AMFF) is proposed. Firstly, the static features of large amounts of unlabeled speech data are input into DAE for unsupervised training. Secondly, the frame-level features and static features of a small amount of labeled speech data are simultaneously input into the LSTM network and the encoded output part of DAE for joint supervised training. Finally, a feature fusion algorithm based on attention mechanism is proposed, which can get the optimal feature set in the training process. Simulation results show that the proposed feature fusion method is significantly better than traditional feature fusion methods, and the model can achieve advanced performance with only a small amount of labeled data.
Yuanbo FANG
Ministry of Education,Henan University of Technology
Hongliang FU
Ministry of Education,Henan University of Technology
Huawei TAO
Ministry of Education,Henan University of Technology,Southeast University
Ruiyu LIANG
Nanjing Institute of Technology
Li ZHAO
Southeast University
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Copy
Yuanbo FANG, Hongliang FU, Huawei TAO, Ruiyu LIANG, Li ZHAO, "A Novel Hybrid Network Model Based on Attentional Multi-Feature Fusion for Deception Detection" in IEICE TRANSACTIONS on Fundamentals,
vol. E104-A, no. 3, pp. 622-626, March 2021, doi: 10.1587/transfun.2020EAL2051.
Abstract: Speech based deception detection using deep learning is one of the technologies to realize a deception detection system with high recognition rate in the future. Multi-network feature extraction technology can effectively improve the recognition performance of the system, but due to the limited labeled data and the lack of effective feature fusion methods, the performance of the network is limited. Based on this, a novel hybrid network model based on attentional multi-feature fusion (HN-AMFF) is proposed. Firstly, the static features of large amounts of unlabeled speech data are input into DAE for unsupervised training. Secondly, the frame-level features and static features of a small amount of labeled speech data are simultaneously input into the LSTM network and the encoded output part of DAE for joint supervised training. Finally, a feature fusion algorithm based on attention mechanism is proposed, which can get the optimal feature set in the training process. Simulation results show that the proposed feature fusion method is significantly better than traditional feature fusion methods, and the model can achieve advanced performance with only a small amount of labeled data.
URL: https://global.ieice.org/en_transactions/fundamentals/10.1587/transfun.2020EAL2051/_p
Copy
@ARTICLE{e104-a_3_622,
author={Yuanbo FANG, Hongliang FU, Huawei TAO, Ruiyu LIANG, Li ZHAO, },
journal={IEICE TRANSACTIONS on Fundamentals},
title={A Novel Hybrid Network Model Based on Attentional Multi-Feature Fusion for Deception Detection},
year={2021},
volume={E104-A},
number={3},
pages={622-626},
abstract={Speech based deception detection using deep learning is one of the technologies to realize a deception detection system with high recognition rate in the future. Multi-network feature extraction technology can effectively improve the recognition performance of the system, but due to the limited labeled data and the lack of effective feature fusion methods, the performance of the network is limited. Based on this, a novel hybrid network model based on attentional multi-feature fusion (HN-AMFF) is proposed. Firstly, the static features of large amounts of unlabeled speech data are input into DAE for unsupervised training. Secondly, the frame-level features and static features of a small amount of labeled speech data are simultaneously input into the LSTM network and the encoded output part of DAE for joint supervised training. Finally, a feature fusion algorithm based on attention mechanism is proposed, which can get the optimal feature set in the training process. Simulation results show that the proposed feature fusion method is significantly better than traditional feature fusion methods, and the model can achieve advanced performance with only a small amount of labeled data.},
keywords={},
doi={10.1587/transfun.2020EAL2051},
ISSN={1745-1337},
month={March},}
Copy
TY - JOUR
TI - A Novel Hybrid Network Model Based on Attentional Multi-Feature Fusion for Deception Detection
T2 - IEICE TRANSACTIONS on Fundamentals
SP - 622
EP - 626
AU - Yuanbo FANG
AU - Hongliang FU
AU - Huawei TAO
AU - Ruiyu LIANG
AU - Li ZHAO
PY - 2021
DO - 10.1587/transfun.2020EAL2051
JO - IEICE TRANSACTIONS on Fundamentals
SN - 1745-1337
VL - E104-A
IS - 3
JA - IEICE TRANSACTIONS on Fundamentals
Y1 - March 2021
AB - Speech based deception detection using deep learning is one of the technologies to realize a deception detection system with high recognition rate in the future. Multi-network feature extraction technology can effectively improve the recognition performance of the system, but due to the limited labeled data and the lack of effective feature fusion methods, the performance of the network is limited. Based on this, a novel hybrid network model based on attentional multi-feature fusion (HN-AMFF) is proposed. Firstly, the static features of large amounts of unlabeled speech data are input into DAE for unsupervised training. Secondly, the frame-level features and static features of a small amount of labeled speech data are simultaneously input into the LSTM network and the encoded output part of DAE for joint supervised training. Finally, a feature fusion algorithm based on attention mechanism is proposed, which can get the optimal feature set in the training process. Simulation results show that the proposed feature fusion method is significantly better than traditional feature fusion methods, and the model can achieve advanced performance with only a small amount of labeled data.
ER -