Multi-Object Tracking in Videos using Advanced Learning Spatio-Temporal Information
The objective of our project is to provide robustness of multi-object tracking. We propose a multi-object tracking
algorithm with temporal-spatial information and trajectory of confidence. The whole process is divided into local and global
association. Trajectories with high confidence are associated with the detection result of the current frame during local
association , whereas trajectories with low confidence are associated with the detection results of the current frame are not
matched during global association. A combined model is to determine association results with help of information of spatial
temporal correlation. Using this model can provide more robust and can also deal with missed detection. The reliability is
measured using the confidence map smoothing constraint and peak side lobe ratio criteria. Our proposed model is more
superior over the many existing algorithms and poor tracker robustness.
Keywords - Multi-Object Tracking, Trajectory of Confidence, Spatio-Temporal Information.