TY - GEN
T1 - An Augmented Treble Stream Deep Neural Network for Video Analysis
AU - Zhang, Chaolong
AU - Xu, Yuanping
AU - Xu, Zhijie
AU - Gong, Mei
AU - Guo, Benjun
AU - Yao, Dengguo
N1 - Publisher Copyright:
© 2020 IEEE.
PY - 2020/9
Y1 - 2020/9
N2 - Video analysis for human action recognition is one of the most important research areas in pattern recognition and computer vision due to its wide applications. Deep learning-based approaches have been proven more effective than conventional feature engineering-based models. However, the performance is still unreliable when facing real-world application scenarios. Inspired by the Convolutional Neural Network (CNN) and Recurrent Long-Short Term Model (LSTM), this paper presents an augmented treble-stream deep neural network architecture that supports direct extraction of spatial-Temporal features from video streams and their corresponding dense optical flows. This innovative approach assists effective detection of complex video event features that are annotated by rich event "appearance"and motion features. Substantially improved recognition accuracy is recorded during the experiments that are carried and benchmarked over public video event datasets, for example, UCF 101 and HMDB 51. Analytical evaluation approves the validity and effectiveness of the treble-stream neural network design.
AB - Video analysis for human action recognition is one of the most important research areas in pattern recognition and computer vision due to its wide applications. Deep learning-based approaches have been proven more effective than conventional feature engineering-based models. However, the performance is still unreliable when facing real-world application scenarios. Inspired by the Convolutional Neural Network (CNN) and Recurrent Long-Short Term Model (LSTM), this paper presents an augmented treble-stream deep neural network architecture that supports direct extraction of spatial-Temporal features from video streams and their corresponding dense optical flows. This innovative approach assists effective detection of complex video event features that are annotated by rich event "appearance"and motion features. Substantially improved recognition accuracy is recorded during the experiments that are carried and benchmarked over public video event datasets, for example, UCF 101 and HMDB 51. Analytical evaluation approves the validity and effectiveness of the treble-stream neural network design.
KW - CNN and LSTM
KW - deep learning
KW - human action recognition
KW - treble-stream network
KW - video analysis
UR - http://www.scopus.com/inward/record.url?scp=85102940003&partnerID=8YFLogxK
U2 - 10.1109/IV51561.2020.00056
DO - 10.1109/IV51561.2020.00056
M3 - Conference Proceeding
AN - SCOPUS:85102940003
T3 - Proceedings of the International Conference on Information Visualisation
SP - 301
EP - 306
BT - 2020 24th International Conference Information Visualisation, IV 2020
A2 - Banissi, Ebad
A2 - Khosrow-Shahi, Farzad
A2 - Ursyn, Anna
A2 - McK. Bannatyne, Mark W.
A2 - Pires, Joao Moura
A2 - Datia, Nuno
A2 - Nazemi, Kawa
A2 - Kovalerchuk, Boris
A2 - Counsell, John
A2 - Agapiou, Andrew
A2 - Vrcelj, Zora
A2 - Chau, Hing-Wah
A2 - Li, Mengbi
A2 - Nagy, Gehan
A2 - Laing, Richard
A2 - Francese, Rita
A2 - Sarfraz, Muhammad
A2 - Bouali, Fatma
A2 - Venturin, Gilles
A2 - Trutschl, Marjan
A2 - Cvek, Urska
A2 - Muller, Heimo
A2 - Nakayama, Minoru
A2 - Temperini, Marco
A2 - Di Mascio, Tania
A2 - Rossano, Filippo Sciarrone Veronica Rossano
A2 - Dorner, Ralf
A2 - Caruccio, Loredana
A2 - Vitiello, Autilia
A2 - Huang, Weidong
A2 - Risi, Michele
A2 - Erra, Ugo
A2 - Andonie, Razvan
A2 - Ahmad, Muhammad Aurangzeb
A2 - Figueiras, Ana
A2 - Mabakane, Mabule Samuel
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 24th International Conference Information Visualisation, IV 2020
Y2 - 7 September 2020 through 11 September 2020
ER -