A Computational Approach to Edge Detection
IEEE Transactions on Pattern Analysis and Machine Intelligence
Simulating humans: computer graphics animation and control
Simulating humans: computer graphics animation and control
Towards model-based recognition of human movements in image sequences
CVGIP: Image Understanding
Digital video analysis and recognition for content-based access
ACM Computing Surveys (CSUR)
3-D model-based tracking of humans in action: a multi-view approach
CVPR '96 Proceedings of the 1996 Conference on Computer Vision and Pattern Recognition (CVPR '96)
Model-based estimation of 3D human motion with occlusion based on active multi-viewpoint selection
CVPR '96 Proceedings of the 1996 Conference on Computer Vision and Pattern Recognition (CVPR '96)
Video as an image data source: efficient representations and applications
ICIP '95 Proceedings of the 1995 International Conference on Image Processing (Vol. 1)-Volume 1 - Volume 1
Content-based video sequence representation
ICIP '95 Proceedings of the 1995 International Conference on Image Processing (Vol.2)-Volume 2 - Volume 2
Tracking human walking in dynamic scenes
ICIP '97 Proceedings of the 1997 International Conference on Image Processing (ICIP '97) 3-Volume Set-Volume 1 - Volume 1
| Hi-index | 0.00 |
In numerous content-based video applications, it isimportant to extract from a video sequence a representation forhumans in motion. This task is difficult, because humans are notrigid objects and they are capable of performing a wide variety ofactions. However, often, human movements can be categorized intorepetitive and rhythmic patterns of motion. Identifying the motionpattern of a human significantly alleviates the task of constructionof its representation. We propose here a model-based recognition ofthe generic posture of human walking in dynamic scenes. We model thehuman body as an articulated object connected by joints and rigidparts, and model the human walking as a periodic motion. Therecognition task is to fit the model walker sequence to the walker inthe live video (data walker sequence). We achieve this by determiningthe period of the data walker sequence and finding its phase withrespect to the model walker sequence. We present promising results ofhow our system performs with a live video sequence.