|
|
Tracking by Spatio-Temporal Segmentation |
|---|---|
| Objective | This project focuses on an advanced tracker that is based on the segmentation of spatio-temporal volumes. |
| Description | In this project we do tracking by segmentation of spatio-temporal volumes. Therefore a video sequence is stacked up in a 3D volume and segmented using a variational energy minimization approach. 
The project will focus on improving the current framework by more accurate descriptors and adding motion estimation (optical flow) to the segmentation framework. Languages and packages used: C++, CUDA, Matlab, Qt |
| Team | Computer Vision |
| Contact | Unger Markus |
| Offered as |
Semester project Master's thesis |
| Duration |
2011-Oct
-
2012-Feb
|
| References |
