Dynamically adaptive correlation is used for robust target tracking.

Proposed algorithm does not require an offline training process.

Prediction and re-detection mechanisms are used to increase tracking efficiency.

Dynamically-adaptive filtering is robust to noise, clutter and illumination.

Algorithm outperforms well-known trackers in terms of objective metrics.