更新:2018-03-12 09:17:28 阅读:人次





Recent Trends in Activity Recognition


Abstract: Twoof our new attempts in combining deep learning with human activity recognitionwill be introduced in this talk. (1) Wi-Fi channel state information (CSI)provides adequate information for recognizing and analyzing human activities.Traditional Wi-Fi CSI based human activity recognition methods adopt Wi-Fi CSIfrom a single AP, which is not so appropriate for a high-density Wi-Fi environment.In this work, we propose a learning method that analyzes the CSI of multipleAPs in a small area to detect and recognize human activities. We introduce adeep learning model to process complex and large CSI information from multipleAPs. From extensive experiment results, our method performs better than othersolutions in a given environment where multiple Wi-Fi APs exist. (2) Thepostures of the same activity performed by different persons may not be thesame exactly. This phenomenon makes it hard to explore a robust activitydescription based on postures. In order to better address the individualdiversity, we propose a skeletal point based concept denoted as ActivityContribution Degree (ACD), which indicates the information contribution levelof the skeletal point. Since the ACD variation of the skeletal point is morestable than postures, we can get a more robust activity description. Theexperimental results show that our model can achieve higher recognitionaccuracy.


Biography: MianxiongDong received B.S., M.S. and Ph.D. in Computer Science and Engineering from TheUniversity of Aizu, Japan. He is currently an Associate Professor in theDepartment of Information and Electronic Engineering at the Muroran Instituteof Technology, Japan. He was a JSPS Research Fellow with School of ComputerScience and Engineering, The University of Aizu, Japan and was a visitingscholar with BBCR group at University of Waterloo, Canada supported by JSPSExcellent Young Researcher Overseas Visit Program from April 2010 to August2011. Dr. Dong was selected as a Foreigner Research Fellow (a total of 3recipients all over Japan) by NEC C&C Foundation in 2011. His researchinterests include Wireless Networks, Cloud Computing, and Cyber-physical Systems.He has received best paper awards from IEEE HPCC 2008, IEEE ICESS 2008, ICA3PP2014, GPC 2015, IEEE DASC 2015, IEEE VTC 2016-Fall, FCST 2017 and 2017 IETCommunications Premium Award. Dr. Dong serves as an Editor for IEEETransactions on Green Communications and Networking (TGCN), IEEE CommunicationsSurveys and Tutorials, IEEE Network, IEEE Wireless Communications Letters, IEEECloud Computing, IEEE Access, as well as a leading guest editor for ACMTransactions on Multimedia Computing, Communications and Applications (TOMM),IEEE Transactions on Emerging Topics in Computing (TETC), IEEE Transactions onComputational Social Systems (TCSS). He has been serving as the Vice Chair ofIEEE Communications Society Asia/Pacific Region Meetings and Conference Committee,Leading Symposium Chair of IEEE ICC 2019, Student Travel Grants Chair of IEEEGLOBECOM 2019, and Symposium Chair of IEEE GLOBECOM 2016, 2017. He is therecipient of IEEE TCSC Early Career Award 2016, IEEE SCSTC Outstanding YoungResearcher Award 2017, The 12th IEEE ComSoc Asia-Pacific Young Researcher Award2017, and Funai Research Award 2018.