DocumentCode
582219
Title
Human activity recognition based on event histogram and KL transform
Author
Jianqin, Yin ; Guohui, Tian ; Xinran, Wang
Author_Institution
Sch. of Control Sci. & Eng., Shandong Univ., Jinan, China
fYear
2012
fDate
25-27 July 2012
Firstpage
3912
Lastpage
3916
Abstract
A human activity representation method based on event histogram was put forward, and the human activity recognition was realized based on event histogram and KL transform. Ubiquitous sensors were used to acquire data and based on these data human recognition was realized. In order to model the sensor data sequences, the concept of event and event histogram was put forward. The event is corresponding to the possible state of the sensor, and the event histogram represents the frequency of the events. At last, event histogram and KL transform were used to realize the human activity recognition. Experimental results show that the method can better realize human activity recognition.
Keywords
Karhunen-Loeve transforms; behavioural sciences; data acquisition; pattern recognition; sensors; KL transform; event histogram; human activity recognition; human activity representation method; sensor data sequences; ubiquitous sensors; Educational institutions; Hidden Markov models; Histograms; Humans; Pattern recognition; Sensors; Transforms; Histogram; Human activity recognition; KL transform;
fLanguage
English
Publisher
ieee
Conference_Titel
Control Conference (CCC), 2012 31st Chinese
Conference_Location
Hefei
ISSN
1934-1768
Print_ISBN
978-1-4673-2581-3
Type
conf
Filename
6390609
Link To Document