IRO Journals
Journal of ISMAC

YOLOv9-based Real-time Human Detection, Tracking, and Directional Counting in Public Environments

Open Access
Published: 17 October, 2026
Pages: 399-410
  • Sesha Phaneendra G. , Sesha Phaneendra G.
    Department of Computer Science and Engineering, Anantha Lakshmi Institute of Technology & Sciences, JNTUA, Ananthapuramu, India
    Department of Computer Science and Engineering, Anantha Lakshmi Institute of Technology & Sciences, JNTUA, Ananthapuramu, India
  • Shaik Affika Sultana Shaik Affika Sultana
    Department of Computer Science and Engineering, Anantha Lakshmi Institute of Technology & Sciences, JNTUA, Ananthapuramu, India
    Department of Computer Science and Engineering, Anantha Lakshmi Institute of Technology & Sciences, JNTUA, Ananthapuramu, India
view PDF
view PDF

How to Cite

G., Sesha Phaneendra, and Shaik Affika Sultana. 2026. “YOLOv9-Based Real-Time Human Detection, Tracking, and Directional Counting in Public Environments”. Journal of ISMAC 8 (4): 399-410. https://doi.org/10.36548/jismac.2026.4.005.
ACM ACS APA ABNT Chicago Harvard IEEE MLA Turabian Vancouver AMA

Download Citation

Endnote/Zotero/Mendeley (RIS) BibTeX

Keywords

YOLOv9
Human Detection
Centroid-Based Tracking
Virtual Line Crossing
Non-Maximum Suppression
CrowdHuman Dataset
Real-time Video Analytics

Abstract

A real-time people-counting system for crowded scenarios needs to have effective human detection, temporal tracking, and estimation of directional movement despite possible occlusions and different scenarios. This research work proposes a real-time people-counting system that utilizes the YOLOv9 algorithm for video-frame processing, Non-Maximum Suppression (NMS), centroid tracking, and virtual line-crossing for the estimation of IN/OUT direction. This study uses the CrowdHuman annotated image dataset, which is split into training, validation, and testing datasets. Human instances detected in the video are localized by bounding boxes and linked with the centroids’ positions between the consecutive frames, and the crossing points are detected by crossing of centroid positions on a certain virtual border. This pipeline can process both pre-recorded videos and live webcam feeds. to the proposed mode has resulted in an accuracy of 96%, 91% precision, and 91% recall.

References

  1. Pore, Satish D., and B. F. Momin. "Bidirectional People Counting System in Video Surveillance." In International Conference on Recent Trends in Electronics, Information Communication Technology (RTEICT), IEEE, 2016, 724-727.
  2. Perng, Jau-Woei, Ting-Yen Wang, Ya-Wen Hsu, and Bing-Fei Wu. "The Design and Implementation of A Vision-Based People Counting System in Buses." In International Conference on System Science and Engineering (ICSSE), IEEE, 2016, 1-3.
  3. Saxena, Surbhi, and Dalpat Songara. "Design of People Counting System using MATLAB." In Tenth International Conference on Contemporary Computing (IC3), IEEE, 2017, 1-3.
  4. Wahyuni, Elvira Sukma, Rizqi Renafasih Alinra, and Hendra Setiawan. "People Counting for Indoor Monitoring." In International Conference on Computing, Engineering, and Design (ICCED), IEEE, 2017, 1-5.
  5. Le, Manh Cuong, My-Ha Le, and Minh-Thien Duong. "Vision-Based People Counting for Attendance Monitoring System." In 5th International Conference on Green Technology and Sustainable Development (GTSD), IEEE, 2020, 349-352.
  6. Holla, M. Raviraja, D. Suma, and M. Darshan Holla. "Optimizing Accuracy and Efficiency in Real-Time People Counting with Cascaded Object Detection." International Journal of Information Technology, 2024, 1-14.
  7. Wang, Chien-Yao, I-Hau Yeh, and Hong-Yuan Mark Liao. "Yolov9: Learning What You Want to Learn Using Programmable Gradient Information." In European conference on computer vision, Cham: Springer Nature Switzerland, 2024, 1-21.
  8. Patwal, Akshita, Manoj Diwakar, Vikas Tripathi, and Prabhishek Singh. "Crowd Counting Analysis using Deep Learning: A Critical Review." Procedia Computer Science, 2023, vol. 218: 2448-2458.
  9. Maktoof, Mohammed Abdul Jaleel. "Crowd Counting using Yolov5 and KCF." Periodicals of Engineering and Natural Sciences, 2023, vol. 11, no. 2: 92-101.
  10. Ahamed, Ishrath, Chamith Dilshan Ranathunga, Dinuka Sandun Udayantha, Benny Kai Kiat Ng, and Chau Yuen. "Real-Time AI-Driven People Tracking and Counting using Overhead Cameras." In TENCON, IEEE, 2024, 952-955.
  11. S. Shao, Z. Zhao, B. Li, T. Xiao, G. Yu, X. Zhang, and J. Sun, “CrowdHuman: A Benchmark for Detecting Human in a Crowd,” IEEE Transactions on Pattern Analysis and Machine Intelligence, 2021. Available: https://www.crowdhuman.org/