Education, Science, Technology, Innovation and Life
Open Access
Sign In
  1. Home
  2. Journals
  3. Journal of Image Processing Theory and Applications
  4. Vol 7, Issue 1, 2024

Research on Pedestrian Localization Methods with Sparse or Unlabeled Occlusion Data

Download as PDF

DOI: 10.23977/jipta.2024.070116 | Downloads: 11 | Views: 698

Author(s)

Xuhao Guo 1, Yongyan Liu 1

Affiliation(s)

1 Department of Artificial Intelligence, University of Melbourne, Melbourne of Victoria, Australia

Corresponding Author

Xuhao Guo

ABSTRACT

In real-world scenarios, sparse or unlabeled occlusion data significantly limits the optimization of pedestrian localization and identification models. This limitation is particularly evident in intelligent urban systems and security applications, where manual labeling is not only costly but also hindered by challenges such as multi-camera coverage and varying degrees of occlusion. The absence of features in occluded regions further exacerbates performance degradation, especially in multi-camera surveillance settings. This study proposes a novel architecture, MaskFormer, based on an Adaptive Masking Mechanism (AMM), which dynamically generates occlusion masks and integrates a local-global feature interaction module to effectively address occlusion recovery and pedestrian feature extraction. Experimental results on the Occluded-Duke dataset demonstrate that MaskFormer significantly outperforms traditional methods in both occlusion recovery performance (PSNR and SSIM) and pedestrian identification metrics (mAP and Rank-1 accuracy), achieving an mAP of 52.8% and a Rank-1 accuracy of 65.4%. Additionally, t-SNE visualization and ablation studies further validate the contributions of each module to the overall performance. The findings not only highlight MaskFormer’s robustness in complex occlusion scenarios but also provide an efficient solution for pedestrian identification tasks in large-scale unlabeled data environments. Future work will focus on enhancing the model's real-time performance, cross-domain generalization capabilities, and integration with multimodal data.

KEYWORDS

Adaptive Masking Mechanism, Local-Global Feature Interaction, Occlusion Region Prediction, Multimodal Data Fusion

CITE THIS PAPER

Xuhao Guo, Yongyan Liu, Research on Pedestrian Localization Methods with Sparse or Unlabeled Occlusion Data. Journal of Image Processing Theory and Applications (2024) Vol. 7: 134-142. DOI: http://dx.doi.org/10.23977/jipta.2024.070116.

REFERENCES

[1] Alfikri M D, Kaliski R. Real-Time Pedestrian Detection on IoT Edge Devices: A Lightweight Deep Learning Approach[J]. arXiv preprint arXiv:2409.15740, 2024.
[2] Deng Guanghong. Research on Pedestrian Detection Methods Based on Deep Learning [D]. Jiangxi University of Science and Technology, 2020.
[3] Xie C, Li P, Sun Y. Pedestrian detection and location algorithm based on deep learning[C]//2019 International Conference on Intelligent Transportation, Big Data & Smart City (ICITBS). IEEE, 2019: 582-585.
[4] Yang M, Huang Z, Hu P, et al. Learning with twin noisy labels for visible-infrared person re-identification[C]// Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. 2022: 14308-14317.
[5] Li Z, Zhang Y, Wang C, et al. Improved Pedestrian Detection Algorithm Based on YOLOv5s[J]. Journal of Advanced Computational Intelligence and Intelligent Informatics, 2024, 28(4): 768-775.
[6] Zhao R, Hao J, Huo H. Research on Multi-Modal Pedestrian Detection and Tracking Algorithm Based on Deep Learning [J]. Future Internet, 2024, 16(6): 194.
[7] Kulhandjian H, Barron J, Tamiyasu M, et al. AI-Based Pedestrian Detection and Avoidance at Night Using Multiple Sensors[J]. Journal of Sensor and Actuator Networks, 2024, 13(3): 34.
[8] Li M, Liu M. Deep-learning-based algorithm for classifying pedestrian behavior at crosswalks[C]//Third International Conference on Image Processing, Object Detection, and Tracking (IPODT 2024). SPIE, 2024, 13396: 102-107.
[9] Bouchamla H, Boumaiza Z, Mabrouk W B, et al. Collision avoidance in pedestrian-rich environments using deep learning[C]//2024 International Conference on Control, Automation and Diagnosis (ICCAD). IEEE, 2024: 1-7.
[10] Sumi A, Santha T. Deep Learning Based Pedestrian Detection using Semantic and Multiscale Deep Features[C]// 2024 10th International Conference on Advanced Computing and Communication Systems (ICACCS). IEEE, 2024, 1: 2100-2105.
[11] Lei S, Yi H, Sarmiento J S. Synchronous End-to-End Vehicle Pedestrian Detection Algorithm Based on Improved YOLOv8 in Complex Scenarios[J]. Sensors, 2024, 24(18): 6116.
[12] Park S, Kim H, Ro Y M. Robust pedestrian detection via constructing versatile pedestrian knowledge bank[J]. Pattern Recognition, 2024, 153: 110539.
[13] Gonthina N, Kola S K, Dabbeti P, et al. Robust Pedestrian Detection in Challenging Environmental Conditions Using FCOS[C]//2023 3rd International Conference on Mobile Networks and Wireless Communications (ICMNWC). IEEE, 2023: 1-6. 

Subscription
E-Mail Alert
Most popular papers
Call for Special Issue
Publication Ethics & OA Statement
Plagiarism Policy
Contact Us
Downloads: 2455
Visits: 172138

Sponsors, Associates, and Links

All published work is licensed under a Creative Commons Attribution 4.0 International License.

Copyright © 2016 - 2031 Clausius Scientific Press Inc. All Rights Reserved.