Learning Long-Term Temporal Features With Deep Neural Networks for Human Action Recognition
One of challenging tasks in the field of artificial intelligence is the human action recognition. In this paper, we propose a novel long-term temporal feature learning architecture for recognizing human action in video, named Pseudo Recurrent Residual Neural Networks (P-RRNNs), which exploits the re...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2020-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/8943218/ |
| _version_ | 1818910068704280576 |
|---|---|
| author | Sheng Yu Li Xie Lin Liu Daoxun Xia |
| author_facet | Sheng Yu Li Xie Lin Liu Daoxun Xia |
| author_sort | Sheng Yu |
| collection | DOAJ |
| description | One of challenging tasks in the field of artificial intelligence is the human action recognition. In this paper, we propose a novel long-term temporal feature learning architecture for recognizing human action in video, named Pseudo Recurrent Residual Neural Networks (P-RRNNs), which exploits the recurrent architecture and composes each in different connection among units. Two-stream CNNs model (GoogLeNet) is employed for extracting local temporal and spatial features respectively. The local spatial and temporal features are then integrated into global long-term temporal features by using our proposed two-stream P-RRNNs. Finally, the Softmax layer fuses the outputs of two-stream P-RRNNs for action recognition. The experimental results on two standard databases UCF101 and HMDB51 demonstrate the outstanding performance of proposed method based on architectures for human action recognition. |
| first_indexed | 2024-12-19T22:36:56Z |
| format | Article |
| id | doaj.art-9c299118491f43df8881593cdace2256 |
| institution | Directory Open Access Journal |
| issn | 2169-3536 |
| language | English |
| last_indexed | 2024-12-19T22:36:56Z |
| publishDate | 2020-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Access |
| spelling | doaj.art-9c299118491f43df8881593cdace22562022-12-21T20:03:10ZengIEEEIEEE Access2169-35362020-01-0181840185010.1109/ACCESS.2019.29622848943218Learning Long-Term Temporal Features With Deep Neural Networks for Human Action RecognitionSheng Yu0https://orcid.org/0000-0001-8302-462XLi Xie1https://orcid.org/0000-0002-3303-312XLin Liu2https://orcid.org/0000-0002-8919-3173Daoxun Xia3School of Information Science and Engineering and Provincial Demonstration Software Institute, Shaoguan University, Shaoguan, ChinaSchool of Information Science and Engineering and Provincial Demonstration Software Institute, Shaoguan University, Shaoguan, ChinaSchool of Information Science and Engineering and Provincial Demonstration Software Institute, Shaoguan University, Shaoguan, ChinaSchool of Big Data and Computer Science, Guizhou Normal University, Guiyang, ChinaOne of challenging tasks in the field of artificial intelligence is the human action recognition. In this paper, we propose a novel long-term temporal feature learning architecture for recognizing human action in video, named Pseudo Recurrent Residual Neural Networks (P-RRNNs), which exploits the recurrent architecture and composes each in different connection among units. Two-stream CNNs model (GoogLeNet) is employed for extracting local temporal and spatial features respectively. The local spatial and temporal features are then integrated into global long-term temporal features by using our proposed two-stream P-RRNNs. Finally, the Softmax layer fuses the outputs of two-stream P-RRNNs for action recognition. The experimental results on two standard databases UCF101 and HMDB51 demonstrate the outstanding performance of proposed method based on architectures for human action recognition.https://ieeexplore.ieee.org/document/8943218/Action recognitionresidual learningrecurrent neural networkslong short-term memory (LSTM) |
| spellingShingle | Sheng Yu Li Xie Lin Liu Daoxun Xia Learning Long-Term Temporal Features With Deep Neural Networks for Human Action Recognition IEEE Access Action recognition residual learning recurrent neural networks long short-term memory (LSTM) |
| title | Learning Long-Term Temporal Features With Deep Neural Networks for Human Action Recognition |
| title_full | Learning Long-Term Temporal Features With Deep Neural Networks for Human Action Recognition |
| title_fullStr | Learning Long-Term Temporal Features With Deep Neural Networks for Human Action Recognition |
| title_full_unstemmed | Learning Long-Term Temporal Features With Deep Neural Networks for Human Action Recognition |
| title_short | Learning Long-Term Temporal Features With Deep Neural Networks for Human Action Recognition |
| title_sort | learning long term temporal features with deep neural networks for human action recognition |
| topic | Action recognition residual learning recurrent neural networks long short-term memory (LSTM) |
| url | https://ieeexplore.ieee.org/document/8943218/ |
| work_keys_str_mv | AT shengyu learninglongtermtemporalfeatureswithdeepneuralnetworksforhumanactionrecognition AT lixie learninglongtermtemporalfeatureswithdeepneuralnetworksforhumanactionrecognition AT linliu learninglongtermtemporalfeatureswithdeepneuralnetworksforhumanactionrecognition AT daoxunxia learninglongtermtemporalfeatureswithdeepneuralnetworksforhumanactionrecognition |