Abstract
This research focuses on the intelligent identification of live behavior and aims to build an efficient and accurate live behavior identification system. The paper first reviews the existing webcast behavior recognition technology and analyzes its advantages and disadvantages. Based on this, an intelligent recognition model of live behavior based on deep learning and multimodal fusion is proposed. The model uses multi-source data such as video, audio, and text to extract features through the deep neural network, and uses an attention mechanism to realize the effective fusion of multi-modal information. To verify the effectiveness of the model, the research team constructed a large-scale data set of webcast behavior, covering a variety of live broadcast scenarios and behavior types. Experimental results show that the proposed model outperforms existing methods in both recognition accuracy and real-time performance. Finally, the model is applied to the actual network broadcast platform, and its practical value in content review and user behavior analysis is discussed. This study provides new ideas and methods for the intelligent identification of network broadcast behavior, which is of great significance for improving the management level and user experience of live broadcast platforms.
References
[1] Li Ming, Wang Qiang, and Zhang Wei. Research progress in Video Behavior Recognition based on Deep Learning. Journal of Computer Science, 2022, 45 (3): 521-544.
[2] Chen Yan, Liu Zhi, and Zheng Yu. Review of multimodal emotion recognition techniques. Journal of Automation, 2023, 49 (2): 201-220.
[3] Zhao Wenbo, Huang Ming, Chen Xiaohong. Real-time voice and emotion recognition method in webcast. Computer Research and Development, 2021, 58 (4): 812-825.
[4] Wang Li, Li Hua, and Zhang Ming. Research on the multimodal feature fusion method based on the attention mechanism. Pattern Recognition and Artificial Intelligence, 2022, 35 (1): 67-79.
[5] Liu Qian, Sun Hong, Zhou Zhihua. Summary of text classification and emotion analysis in live broadcast scenes. Journal of Software, 2023, 34 (3): 1-20.
[6] Zhang S, Li S, and Wang W. Research on the application of transfer learning in webcast Behavior Identification. Chinese Science: Information Science, 2021, 51 (8): 1289-1305.
[7] Daniel W, Edison C, and Andy L. Multimodal feature fusion method based on graph neural network. Journal of Computer-Aided Design and Graphics, 2022, 34 (5): 789-801.
[8] Lin ZL, Fan BB, Yang M. Research on real-time human pose estimation technology under live broadcast scene. Chinese Journal of Graphic Pictures, 2023, 28 (2): 345-358.
[9] J Chou, JJ Lin, and LH Wang. Multi-task learning method in webcast content review. Journal of Communications, 2022, 43 (7): 112-126.
How to cite this paper
Research on the Construction and Practice of Intelligent Identification of Webcast Behavior
How to cite this paper: Feng Chen. (2024) Research on the Construction and Practice of Intelligent Identification of Webcast Behavior. Advances in Computer and Communication, 5(3), 200-204.
DOI: http://dx.doi.org/10.26855/acc.2024.07.008