Real-time 3D model-based motion tracking
A new model-based tracking algorithm is reported for real-time motion tracking per-formance. In this experiment, a single calibrated camera and geometric models are used. Each tracking cycle completes computation between a projected 3D model and the image. The early stage is to match and measure the...
Saved in:
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Theses and Dissertations |
| Language: | English |
| Published: |
2010
|
| Subjects: | |
| Online Access: | http://hdl.handle.net/10356/38998 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Nanyang Technological University |
| Language: | English |
| Summary: | A new model-based tracking algorithm is reported for real-time motion tracking per-formance. In this experiment, a single calibrated camera and geometric models are used. Each tracking cycle completes computation between a projected 3D model and the image. The early stage is to match and measure the error between the model and corresponding image features; The later stage is to estimate the model pose according to previous measurement. The matching process includes two aspects of:(l)feature extraction using local minimum energy and (2)global matching of a known three-dimensional model against the projected features. A one-dimensional search strategy is adopted instead of local window operation. The feature energy which is defined as the negative absolute value of the edge strength, fits the small motion hypothesis. The algorithm reported here is robust to change in lighting and backgrounds. |
|---|