|
Rectangular layer model for profile-based human action recognition using multi-view depth information |
|---|---|
| รหัสดีโอไอ | |
| Creator | Pongsagorn Chalearnnetkul |
| Title | Rectangular layer model for profile-based human action recognition using multi-view depth information |
| Contributor | Nikom Suvonvorn |
| Publisher | Research and Technology Transfer Affairs Division.Khon Kaen University. |
| Publication Year | 2560 |
| Journal Title | Asia-Pacific Journal of Science and Technology (APST) |
| Journal Vol. | 22 |
| Journal No. | 3 |
| Page no. | 1-6 |
| Keyword | action recognition, multi-view, depth information, rectangular layer model |
| URL Website | https://tci-thaijo.org/index.php/APST/index |
| Website title | https://tci-thaijo.org/index.php/APST/article/view/101489 |
| ISSN | 2539-6293 |
| Abstract | Human action recognition is a fundamental step for understanding complex activitiesor behaviors, especially for video surveillance and health-care applications. In this paper,we introduce a profile-based human action recognition model based on data frommulti-view cameras and using RGB-D information, which we call theRectangular LayerModel (RLM). Our modelwas able toimprove performance over the single-view approachby ameliorating perspective distortion and providing moreinformation. Thisfusion model was tested for five basic actions: walking and standing, sitting, bending, and lying down, at different perspective viewpoints. The system was able toperform at 28.99 fps,andits overall precision is significant at about 92.25%. |
