Abstract | ||
---|---|---|
In this paper, a novel approach of data field is proposed to discover the action pattern of real-time person tracking, and potential function is presented to find out the power of a person with suspicious action. Firstly, a data field on the first feature is used to find the individual attributes, associated with the velocity, direction changing frequency and appearance frequency respectively. Secondly, the common characteristic of each attribute is obtained by the data field on the main feature from the data field created before. Thirdly, the weighted Euclidean distance classifier is used to identify whether a person is a suspect or not. Finally, the results of the experiment show that the proposed way is feasible and effective in action mining. |
Year | DOI | Venue |
---|---|---|
2008 | 10.1007/978-3-540-88192-6_80 | ADMA |
Keywords | Field | DocType |
main feature,common characteristic,action pattern,real-time person tracking,individual attribute,experiment show,data field,appearance frequency,suspicious action,action mining,real time,euclidean distance | Data field,Data mining,Computer vision,Computer science,Euclidean distance,Artificial intelligence,Suspect,Classifier (linguistics),Machine learning | Conference |
Volume | ISSN | Citations |
5139 | 0302-9743 | 0 |
PageRank | References | Authors |
0.34 | 5 | 4 |
Name | Order | Citations | PageRank |
---|---|---|---|
Shuliang Wang | 1 | 202 | 32.66 |
Juebo Wu | 2 | 14 | 3.27 |
Feng Cheng | 3 | 0 | 0.34 |
Hong Jin | 4 | 0 | 0.34 |