Using YOLOv7 Combined with Visual Attention Mechanism to Achieve Real-time Tracking of Passenger Gaze Targets in Urban Rail Interactive Guidance Interface
Main Article Content
Abstract
Real-time gaze target tracking in urban rail interactive guidance interfaces faces significant challenges due to multitarget occlusion, limited computational resources, and the need for low-latency visual perception. This study proposes a lightweight tracking framework that integrates an improved YOLOv7 detector with the Convolutional Block Attention Module (CBAM) and an enhanced ByteTrack algorithm to achieve efficient and accurate passenger gaze target localization. MobileNetV3 is adopted as the backbone network to reduce computational complexity, while BiFPN and Gaussian Soft-NMS improve multi-scale feature fusion and detection robustness. MediaPipe and PnP-based gaze estimation are combined with an NSA Kalman filter and GIoU matching strategy to enhance temporal consistency and tracking stability under complex scenarios. Experimental results demonstrate that the proposed framework achieves a mAP@0.5 of 96.3%, a MOTP of 85.7%, and 32.2 FPS on embedded platforms, maintaining reliable performance under severe occlusion conditions. Beyond intelligent transportation applications, the proposed lightweight visual perception and attention modeling framework provides useful methodological references for adaptive sensing, edge intelligence, and real-time signal interpretation in electromagnetic wave propagation environments and antenna-assisted human– machine interaction systems.
Downloads
Article Details
This work is licensed under a Creative Commons Attribution 4.0 International License.
Authors who publish with this journal agree to the following terms:
- Authors retain copyright and grant the journal right of first publication with the work simultaneously licensed under a Creative Commons Attribution License that allows others to share the work with an acknowledgement of the work's authorship and initial publication in this journal.
- Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgement of its initial publication in this journal.
- Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work (See The Effect of Open Access).
References
K. Wang, C. Shen, M. Li, and J. Li, “Research on Users’ Willingness to Use the Urban Subway Wayfinding Signage System Based on the DeLone & McLean Model Theory: A Case Study of Wuxi Subway,” Systems, vol. 12, no. 12, pp. 529-556, 2024, doi: 10.3390/systems12120529.
X. Li, Y. Zhang, D. He, X. Teng, B. Liu, and Y. Chen, “Passenger flow detection in subway stations based on improved you only look once algorithm,” Transportation research record, vol. 2677, no. 9, pp. 397-409, 2023, doi: 10.1177/03611981231159128.
U. Junyan W, U. Xia L I, and I. Yazhuo L, “Application Research on Passenger Flow Real-time Detection of Subway Station Platform Based on Deep,” Journal of Jianghan University (Natural Science Edition), vol. 51, no. 5, pp. 67-75, 2023.
W. Zhang, C. Zhu, Y. Qu, G. Liu, and D. H. Lee, “An M-DeepSORT Algorithm for Pedestrian Detection and Tracking Based on Video Images – A Case Study in Ji-nan Subway Station,” Promet-Traffic&Transportation, vol. 37, no. 2, pp. 338-360, 2025, doi: 10.7307/ptt.v37i2.635.
M. Sakthi S, M. Ranjitha T, and M. Vidhya Shree, “Real-Time Passenger Counting and Occupancy-Based Announcement System for Metro Trains Using YOLO-Based Deep Learning,” Journal of Machine Learning and Signal Processing, vol. 1, no. 1, pp. 1-11, 2025.
N. Li, X. Bai, X. Shen, P. Xin, J. Tian, T. Chai, et al., “Dense pedestrian detection based on GR-YOLO,” Sensors, vol. 24, no. 14, pp. 4747-4765, 2024, doi: 10.3390/s24144747.
H. Wang, H. Pei, and J. Zhang, “Detection of locomotive signal lights and pedestrians on railway tracks using improved YOLOv4,” IEEE Access, vol. 10, no. 1, pp. 15495-15505, 2022, doi: 10.1109/ACCESS.2022.3148182.
J. Zuo, Y. Wang, and R. Wang, “Method for Acquiring Passenger Standing Positions on Subway Platforms Based on Binocular Vision,” IEEE Access, vol. 13, no. 1, pp. 32971-32980, 2025, doi: 10.1109/ACCESS.2025.3543130.
G. Zidani, D. Djarah, A. Benmakhlouf, and L. Khettache, “Optimizing Pedestrian Tracking for Robust Perception With Yolov8 and Deepsort,” Applied Computer Science, vol. 20, no. 1, pp. 72-84, 2024, doi: 10.35784/acs-2024-05.
L. Li, F. Gao, S. Ling, Z. Guo, J. Zuo, M. Goodsite, et al., “Would you like to get on the bus? An eye-tracking study based on the stimulus-organism-response framework,” Transportation Research Part F: Traffic Psychology and Behaviour, vol. 109, no. 1, pp. 1114-1136, 2025, doi: 10.1016/j.trf.2025.01.014.
Z. Zeng, K. Zhang, and B. Zhang, “Study on the influence of spatial attributes on passengers’ path selection at fengtai highspeed railway station based on eye tracking,” Buildings, vol. 14, no. 9, pp. 3012-3030, 2024, doi: 10.3390/buildings14093012.
K. Kim J, J. Park, K. Moon Y, and J. Kang S, “Improving gaze tracking in large screens with symmetric gaze angle amplification and optimization technique,” IEEE Access, vol. 11, no. 1, pp. 85799-85811, 2023, doi: 10.1109/ACCESS.2023.3282185.
F. Tang, F. Yang, and X. Tian, “Long-distance person detection based on YOLOv7,” Electronics, vol. 12, no. 6, pp. 1502-1515, 2023, doi: 10.3390/electronics12061502.
C. Li, Y. Wang, and X. Liu, “An improved YOLOv7 lightweight detection algorithm for obscured pedestrians,” Sensors, vol. 23, no. 13, pp. 5912-5925, 2023, doi: 10.3390/s23135912.
J. Hu, Y. Zhou, H. Wang, P. Qiao, and W. Wan, “Research on Deep Learning Detection Model for Pedestrian Objects in Complex Scenes Based on Improved YOLOv7,” Sensors, vol. 24, no. 21, pp. 6922-6947, 2024, doi: 10.3390/s24216922.
G. Garcia, A. Velastin S, N. Lastra, H. Ramirez, S. Seriani, and G. Farias, “Train station pedestrian monitoring pilot study using an artificial intelligence approach,” Sensors, vol. 24, no. 11, pp. 3377-3397, 2024, doi: 10.3390/s24113377.
S. Ennaama, H. Silkan, A. Bentajer, and A. Tahiri, “Enhanced real-time object detection using YOLOv7 and MobileNetv3,” Engineering, Technology & Applied Science Research, vol. 15, no. 1, pp. 19181-19187, 2025, doi: 10.48084/etasr.8777.
C. Liu, Y. Li, J. Gu, Y. Lou, and T. Shen, “Enhancing Pedestrian Target Recognition in Open Community Multi-Scene Spaces Using the Yolo-Stp Network,” ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences, vol. 10, no. 1, pp. 319-328, 2023, doi: 10.5194/isprs-annals-X-1-W1-2023-319-2023.
Y. Hao and C. Geng, “Improved pedestrian vehicle detection for small objects based on attention mechanism,” Int. J. Adv. Netw. Monit. Control, vol. 9, no. 1, pp. 80-89, 2024, doi: 10.2478/ijanmc-2024-0030.
A. Assefa A, W. Tian, N. Acheampong K, M. U. Aftab, and M. Ahmad, “Small-Scale and Occluded Pedestrian Detection Using Multi Mapping Feature Extraction Function and Modified Soft-NMS,” Computational intelligence and neuroscience, vol. 2022, no. 1, pp. 1-11, 2022, doi: 10.1155/2022/9325803.
H. Khaleel A, H. Abbas T, and W. Sami Ibrahim A, “Best low-cost methods for real-time detection of the eye and gaze tracking,” i-com, vol. 23, no. 1, pp. 79-94, 2024, doi: 10.1515/icom-2023-0026.
C. Leblond-Menard and S. Achiche, “Non-intrusive real time eye tracking using facial alignment for assistive technologies,” IEEE Transactions on Neural Systems and Rehabilitation Engineering, vol. 31, no. 1, pp. 954-961, 2023, doi: 10.1109/TNSRE.2023.3236886.
S. Zhuang, Z. Zhao, L. Cao, D. Wang, C. Fu, and K. Du, “A robust and fast method to the perspective-n-point problem for camera pose estimation,” IEEE Sensors Journal, vol. 23, no. 11, pp. 11892-11906, 2023, doi: 10.1109/JSEN.2023.3266392.
P. Roch, B. Shahbaz Nejad, M. Handte, and P. J. Marron, “Axes-aligned non-linear optimized PnP algorithm,” Machine Vision and Applications, vol. 35, no. 6, pp. 137-150, 2024, doi: 10.1007/s00138-024-01618-z.
L. You, Y. Chen, C. Xiao, C. Sun, and R. Li, “Multi-object vehicle detection and tracking algorithm based on improved YOLOv8 and ByteTrack,” Electronics, vol. 13, no. 15, pp. 3033-3048, 2024, doi: 10.3390/electronics13153033.
G. Jang J and W. Kwon Y, “Trajectory Similarity-Based Traffic Flow Analysis Using YOLO+ ByteTrack,” Journal of Multimedia Information System, vol. 12, no. 1, pp. 27-40, 2025, doi: 10.33851/JMIS.2025.12.1.27.
Q. Wang, G. Ye, Q. Chen, S. Zhang, and F. Wang, “A UAV perspective based lightweight target detection and tracking algorithm for intelligent transportation,” Complex & Intelligent Systems, vol. 11, no. 1, pp. 73-90, 2025, doi: 10.1007/s40747-024-01687-7.
N. Zelin R E, P. Lan, W. Chao, L. Jiaheng, and Z. Fangyan, “Low-light pedestrian detection and tracking algorithm based on autoencoder structure and improved Bytetrack,” Journal of Applied Optics, vol. 45, no. 3, pp. 616-629, 2024, doi: 10.5768/JAO202445.0302001.