Far East Journal of Experimental and Theoretical Artificial Intelligence
Volume 1, Issue 1, Pages 45 - 60
(February 2008)
|
|
INTELLIGENT RECOGNITION SCHEME FOR HUMAN ACTIVITIES BY USING 2D HETEROGENEOUS MULTIMODAL SENSATIONS AND INFORMATION TRANSCODING
I-Horng Jeng (Taiwan), Chih-Jen Lee (Taiwan), Tai-Ning Yang (Taiwan) and Allen Y. Chang (Taiwan)
|
Abstract:
In the ambient intelligence era, human activity recognition should be the key technology that makes human computer interaction and brings them into an intelligent, secure and comfortable life style. The intelligent recognition scheme presented in this paper tends to simultaneously satisfy the surveillants and people under surveillance by recognizing the information of their activities. As people debate whether privacy and security are in conflict, requiring trade-offs between the two, or if privacy can enhance security, we try to propose an intelligent scheme here to enhance security while maintaining their privacies.
The ambient intelligence extends the territory from office building to home space for security, automation, and whenever human needs better daily lives. In this paper, we propose a method called 2D-heterogeneous multimodal sensation which tends to fulfill the requirement in ambient intelligence but hiding the privacy. Two modes of information in distance and temperature are used to make sensation fusion then detect the position, postures and activities of human being. Finally, three types of transcoding sequences are selected as candidates to transfer and store the sensed information in private and safety. |
| Keywords and phrases: ambient intelligence, human activity recognition, human computer interaction, intelligent surveillance, privacy and security. |
| Communicated by Shun-Feng Su |
| Number of Downloads: 141 | Number of Views: 630 |
|
