Multi-stream Deep Networks for Human Action Recognition with Sequential Tensor Decomposition

	Multi-stream Deep Networks for Human Action Recognition with Sequential Tensor Decomposition
	Guo, Huiwen; Wu, Xinyu; Feng Wei
刊名	Signal Processing
	2017
文献子类	期刊论文
英文摘要	Effective spatial-temporal representation of motion information is crucial to human action classification. In spite of the attempt of most existing methods capturing spatial-temporal structure and learning motion representations with deep neural networks, such representations are failing to model action at their full temporal extent. To address this problem, this paper proposes a global motion representation by using sequential low-rank tensor decomposition. Specifically, we model an action sequence as a third-order tensor with spatiotemporal structure. Then, by using low-rank tensor decomposition, partial motion of objects in global context were preserved which will be feeding into deep architecture to automatically learning global-term motion features. To simultaneously exploit static spatial features, short-term motion and global-term motion in the video, we describe a multi-stream framework with recurrent convolutional architectures which is end-to-end trainable. Gated Recurrent Unit (GRU) is used as our recurrent unit which have fewer parameters than Long Short-Term Memory (LSTM). Extensive experiments were conducted on two challenging dataset: HMDB51 and UCF101. Experimental results show that our method achieves state-of-the-art performance on the HMDB51 dataset, and is comparable to the state-of-the-art methods on the UCF101 dataset.
URL标识	查看原文
语种	英语
内容类型	期刊论文
源URL	[http://ir.siat.ac.cn:8080/handle/172644/11717]
专题	深圳先进技术研究院_集成所
作者单位	Signal Processing
推荐引用方式 GB/T 7714	Guo, Huiwen,Wu, Xinyu,Feng Wei. Multi-stream Deep Networks for Human Action Recognition with Sequential Tensor Decomposition[J]. Signal Processing,2017.
APA	Guo, Huiwen,Wu, Xinyu,&Feng Wei.(2017).Multi-stream Deep Networks for Human Action Recognition with Sequential Tensor Decomposition.Signal Processing.
MLA	Guo, Huiwen,et al."Multi-stream Deep Networks for Human Action Recognition with Sequential Tensor Decomposition".Signal Processing (2017).