收藏切换
Neural Architecture Search on Temporal Convolutions for Complex Action Recognition
收藏切换
Pengzhen Ren1, 2, Xiaodan Liang1, 3, Xiaojun Chang4, Ziying Zhao1, Yun Xiao5, *
Journal of Computer Research and Development | 2025, 62(8) : 1862 - 1874
Less
收藏切换
Journal of Computer Research and Development | 2025, 62(8): 1862-1874
• Artificial Intelligence •
Neural Architecture Search on Temporal Convolutions for Complex Action Recognition
Full
Pengzhen Ren1, 2, Xiaodan Liang1, 3, Xiaojun Chang4, Ziying Zhao1, Yun Xiao5, *
Affiliations
  • 1Pengcheng Laboratory, Shenzhen, Guangdong 518000
  • 2DarkMatter AI Research, Guangzhou 511466
  • 3School of Intelligent Engineering, Sun Yat-sen University, Shenzhen, Guangdong 518107
  • 4School of Information Science and Technology, University of Science and Technology of China, Hefei 230026
  • 5School of Information Science and Technology, Northwest University, Xi’an 710127
  • born in 1993. PhD, engineer. His main research interests include multi-modal representation learning, visual language pre-training, and model automation design

    born in 1991. PhD, professor, PhD supervisor. Member of CCF. Her main research interests include computer vision, natural language understanding, and smart driving

    born in 1986. PhD, professor, PhD supervisor. His main research interests include multi-modal learning, computer vision, and green artificial intelligence

    born in 1985. PhD, senior engineer, PhD supervisor. Her main research interests include artificial intelligence and spatial analysis, and the application of large model technology

    born in 1978. PhD, professor, PhD supervisor. Senior member of CCF. Her main research interests include data mining, machine learning, and artificial intelligence algorithm research and applications

About Author:

born in 1993. PhD, engineer. His main research interests include multi-modal representation learning, visual language pre-training, and model automation design

born in 1991. PhD, professor, PhD supervisor. Member of CCF. Her main research interests include computer vision, natural language understanding, and smart driving

born in 1986. PhD, professor, PhD supervisor. His main research interests include multi-modal learning, computer vision, and green artificial intelligence

born in 1985. PhD, senior engineer, PhD supervisor. Her main research interests include artificial intelligence and spatial analysis, and the application of large model technology

born in 1978. PhD, professor, PhD supervisor. Senior member of CCF. Her main research interests include data mining, machine learning, and artificial intelligence algorithm research and applications

Published: 2025-08-01 doi: 10.7544/issn1000-1239.202440048
Outline
收藏切换

In the field of complex action recognition in videos, the structural design of the model plays a crucial role in its final performance. However, manually designed network structures often rely heavily on the knowledge and experience of researchers. Therefore, neural architecture search (NAS) has received widespread attention from researchers in the field of image processing because of its automated network structure design. Currently, neural architecture search has achieved tremendous development in the image field. Some NAS methods even reduce the number of graphics processing unit (GPU) days required for automated model design to single digits, and the model structures they search show strong competitive potential. This encourages us to extend automated model structure design to the video domain. But it faces two serious challenges: 1) How to capture the long-range contextual temporal association in video as much as possible; 2) How to reduce the computational surge caused by 3D convolution as much as possible. To address the above challenges, we propose a novel neural architecture search on temporal convolutions for complex action recognition (NAS-TC). NAS-TC is a two-stage framework. In the first stage, we use the classic convolutional neural network (CNN) network as the backbone network to complete the computationally intensive feature extraction task. In the second stage, we propose a neural architecture search layer temporal convolutional layer (NAS-TC) to accomplish relatively lightweight long-range temporal model design and information extraction. This ensures that our method will have a more reasonable parameter allocation and can handle minute-level videos. Finally, the proposed method achieves an average performance gain of 2.3% mAP on three complex action recognition benchmark data sets compared with similar methods, and the number of parameters is reduced by 28.5%.

complex action recognition  /  neural architecture search  /  convolution decomposition  /  video understanding  /  deep learning
任鹏真, 梁小丹, 常晓军, 赵子莹, 肖云. 基于时间卷积神经架构搜索的复杂动作识别. 计算机研究与发展, 2025 , 62 (8) : 1862 -1874 . DOI: 10.7544/issn1000-1239.202440048
Pengzhen Ren, Xiaodan Liang, Xiaojun Chang, Ziying Zhao, Yun Xiao. Neural Architecture Search on Temporal Convolutions for Complex Action Recognition[J]. Journal of Computer Research and Development, 2025 , 62 (8) : 1862 -1874 . DOI: 10.7544/issn1000-1239.202440048
Year 2025 volume 62 Issue 8
79
34
Cite this Article
BibTeX
Article Info
doi: 10.7544/issn1000-1239.202440048
  • Receive Date:2024-01-29
  • Online Date:2026-11-10
  • Published:2025-08-01
Article Data
Affiliations
History
  • Received:2024-01-29
  • Revised:2025-02-25
  • Accepted:2025-03-03
Affiliations
    1Pengcheng Laboratory, Shenzhen, Guangdong 518000
    2DarkMatter AI Research, Guangzhou 511466
    3School of Intelligent Engineering, Sun Yat-sen University, Shenzhen, Guangdong 518107
    4School of Information Science and Technology, University of Science and Technology of China, Hefei 230026
    5School of Information Science and Technology, Northwest University, Xi’an 710127
References
Share
https://castjournals.cast.org.cn/joweb/jsjyjyfz/EN/10.7544/issn1000-1239.202440048
Share to
QR

Scan QR to access full text

Cite this article
BibTeX
Citations
表12种不同金属材料的力学参数
科
Family
属数
Number of
genus
种数
Number of
species
占总种数比例
Percentage of
total species (%)
属
Genus
种数
Number of
species
占总种数比例
Percentage of total
species (%)
鹅膏菌科Amanitaceae 2 11 5.26 鹅膏菌属 Amanita 10 4.78
小菇科 Mycenaceae 2 12 5.74 丝盖伞属 Inocybe 5 2.39
多孔菌科 Polyporaceae 8 14 6.70 蜡蘑属 Laccaria 5 2.39
红菇科 Russulaceae 3 23 11.00 小皮伞属 Marasmius 6 2.87
小菇属 Mycena 11 5.26
光柄菇属 Pluteus 5 2.39
红菇属 Russula 17 8.13
栓菌属 Trametes 5 2.39
关闭全屏
  • BibTeX
  • EndNote
  • RefWorks
  • TxT