Enhanced Pedestrian Tracking using Grey Wolf Optimizer

Abstract

Object tracking is a challenging task due to variations in lighting, occlusions, and interactions with other objects, which can impact the appearance model and cause tracking failures. Accurate feature selection is critical for building robust models that can handle noise and sudden changes without drifting. This study introduces an intelligent tracking algorithm, GWO-OF (Grey Wolf Optimizer with Optical Flow), which leverages the hierarchical leadership and hunting behavior of grey wolves for feature optimization and integrates optical flow for motion tracking. The proposed method was evaluated using standard tracking metrics, achieving superior results across all benchmarks. It recorded the highest accuracy (MOTP: 80.53%), precision (MOTA: 48.35%), and F1-Score (73.54%), outperforming state-of-the-art methods such as Memetic-Adaboost and SVM. The GWO-OF algorithm demonstrated robustness in handling occlusions and noise, making it highly effective for real-world tracking applications. These findings highlight the potential of the proposed method as a reliable and precise solution for object tracking.

References
[1] Kokul, Thanikasalam, Amirthalingam Ramanan, and U. A. J. Pinidiyaarachchi. “Online multi-person tracking-by-detection method using ACF and particle filter.” In 2015 IEEE Seventh International Conference on Intelligent Computing and Information Systems (ICICIS), pp. 529-536. IEEE, 2015.
[2] Mantini, Pranav, and Shishir K. Shah. “Multiple people tracking using contextual trajectory forecasting.” In 2016 IEEE Symposium on Technologies for Homeland Security (HST), pp. 1-6. IEEE, 2016.
[3] Shaamili, R., and A. Ruhan Bevi. “An Improved Parallel Feature Fusion-CBAM method for occluded pedestrian identification.” (2023).
[4] Yang, Bin, Junjie Yan, Zhen Lei, and Stan Z. Li. “Aggregate channel features for multi-view face detection.” In IEEE international joint conference on biometrics, pp. 1-8. IEEE, 2014.
[5] Kolluri, Johnson, and Ranjita Das. "Intelligent multimodal pedestrian detection using hybrid metaheuristic optimization with deep learning model." Image and Vision Computing (2023): 104628
[6] Shen, Zheng, Wenxiang Wu, Shuaiya Sun, and Kan Hu. "A Pedestrian Detection Method Based on Improved ResNet for Container Terminal." In Human Centered Computing: 7th International Conference, HCC 2021, Virtual Event, December 9–11, 2021, Revised Selected Papers, pp. 119-132. Cham: Springer Nature Switzerland, 2023.
[7] Qi, Zhouming, Mian Zhou, Guoqiang Zhu, and Yanbing Xue. "Multiple Pedestrian Tracking in Dense Crowds Combined with Head Tracking." Applied Sciences 13, no. 1 (2023): 440.
[8] Praveenkumar, S. M., Prakashgouda Patil, and P. S. Hiremath. "Real-time multi-object tracking of pedestrians in a video using convolution neural network and Deep SORT." In ICT Systems and Sustainability: Proceedings of ICT4SD 2021, Volume 1, pp. 725-736. Springer Singapore, 2022
[9] Mirjalili, Seyedali, Seyed Mohammad Mirjalili, and Andrew Lewis. "Grey wolf optimizer." Advances in engineering software 69 (2014): 46-61.
[10] https://www.vision.caltech.edu/Image_Datasets/CaltechPedestrians.