| |
|
Description: |
Visual feature-based tracking systems need to adapt to variations in the appearance of an object and in the scene for robust performance to avoid drift. Tracking in aerial imagery is challenging as viewing geometry, calibration inaccuracies, complex flight paths and back- ground changes combined with illumination changes, and occlusions can result in rapid appearance change of objects. Balancing appearance adaptation with stability to avoid tracking non-target objects can lead to longer tracks which is an indicator of tracker robustness. We use a Radon transform method with multiscale Hessian for orientation estimation. To handle occlusions in wide-area motion imagery we use a robust partial model approach and rich set of features with maximum likelihood estimation. For robust motion analysis we developed the Flux Tensor with Split Gaussian models (FTSG) that exploits the benefits of fusing low level motion processing method based on spatio-temporal tensor formulation, with a novel foreground and background modeling scheme, and a multi-cue appearance model. This hybrid system can handle challenges such as shadows, illumination changes, dynamic background, stopped and removed objects. Extensive testing performed on the CVPR 2014 Change Detection benchmark dataset shows that FTSG outperforms state-of-the-art methods.
|
|
Date: |
|
| Start Time: |
11:00 |
|
Speaker: |
Kannappan Palaniappan (Univ. Missouri, Columbia, USA)
|
|
Institution: |
Univ. Missouri, Columbia, USA
|
|
Place: |
Room 5.5
|
| Research Groups: |
-Analysis
-Numerical Analysis and Optimization
|
|
See more:
|
<Main>
|
|
