QQ客服 官方微博 反馈留言

	Real-time Action Recognition with Enhanced Motion Vector CNNs
	Bowen Zhang; Limin Wang; Zhe Wang; Yu Qiao; Hanli Wang
	2016
会议名称	CVPR2016
会议地点	美国
英文摘要	The deep two-stream architecture [23] exhibited excellent performance on video based action recognition. The most computationally expensive step in this approach comes from the calculation of optical flow which prevents it to be real-time. This paper accelerates this architecture by replacing optical flow with motion vector which can be obtained directly from compressed videos without extra calculation. However, motion vector lacks fine structures, and contains noisy and inaccurate motion patterns, leading to the evident degradation of recognition performance. Our key insight for relieving this problem is that optical flow and motion vector are inherent correlated. Transferring the knowledge learned with optical flow CNN to motion vector CNN can significantly boost the performance of the latter. Specifically, we introduce three strategies for this, initialization transfer, supervision transfer and their combination. Experimental results show that our method achieves comparable recognition performance to the state-of-the-art, while our method can process 390.7 frames per second, which is 27 times faster than the original two-stream method.
收录类别	EI
语种	英语
内容类型	会议论文
源URL	[http://ir.siat.ac.cn:8080/handle/172644/10008]
专题	深圳先进技术研究院_集成所
作者单位	2016
推荐引用方式 GB/T 7714	Bowen Zhang,Limin Wang,Zhe Wang,et al. Real-time Action Recognition with Enhanced Motion Vector CNNs[C]. 见:CVPR2016. 美国.

个性服务

查看访问统计

相关权益政策

暂无数据

收藏/分享

除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。