Efficient Object Tracking through Machine Learning-Based Meta-Learning

Strategic Insights -

New research from Jadavpur University introduces a machine learning-based approach to one-shot meta-learning for more efficient object tracking. The proposed hybrid model refines predictions from traditional tracking methods using machine learning to enhance performance, achieving a frame rate of 13.74 FPS while maintaining a trade-off between accuracy and computational efficiency. This method offers lower computational complexity, requires fewer resources, and needs minimal training data, making it suitable for applications with moderate latency tolerance.

Key Takeaways:

  • The proposed hybrid model combines traditional tracking methods with machine learning to enhance performance, offering lower computational complexity and fewer resources.
  • The model achieves a frame rate of 13.74 FPS, with a trade-off between accuracy and computational efficiency.
  • The meta-learning-based approach outperforms existing deep-learning models on the VOT2017, OTB50, and GOT10K datasets in most cases.
  • On the OTB50 dataset, the model with XGBoost achieved an OTE of 9.12%, a precision of 92.0%, and a success rate of 86.0%.
  • On the GOT10k dataset, the model with a Decision Tree meta-learner recorded an average overlap of 62.2%, with success rates of 57.9% at 0.5 IoU and 52.0% at 0.75 IoU.
  • On the VoT dataset, the Decision Tree meta-learner attained 79.0% EAO, 88.0% accuracy, and a robustness score of 20.0%.
  • The proposed model is suitable for applications with moderate latency tolerance, making it a viable option for real-world object tracking scenarios.

Statistics:

  • Frame rate: 13.74 FPS
  • OTB50 dataset results:

+ OTE: 9.12%

+ Precision: 92.0%

+ Success rate: 86.0%

  • GOT10k dataset results:

+ Average overlap: 62.2%

+ Success rates at 0.5 IoU: 57.9%

+ Success rates at 0.75 IoU: 52.0%

  • VoT dataset results:

+ EAO: 79.0%

+ Accuracy: 88.0%

+ Robustness score: 20.0%

Sources:

  • MMLT: Efficient object tracking through machine learning-based meta-learning. Results in Engineering, 2025,26():104768. (Results in Engineering - https://www.journals.elsevier.com/results-in-engineering)
  • Elsevier. Results in Engineering. [Online] Available: https://www.journals.elsevier.com/results-in-engineering
  • NewsRx. Findings from Jadavpur University Provide New Insights into Machine Learning (MMLT: Efficient object tracking through machine learning-based meta-learning). Journal of Engineering. June 9, 2025; p 842.