Multi-Traffic Scene Perception

Thireesha Suryadevara; Mudassir Rafi; Nandini Mokhamatam; Sai Keerthi Aluri; Vishnu Priya; Harika Kommu

doi:10.2991/978-94-6239-707-1_14

<Previous Article In Volume

Next Article In Volume>

Multi-Traffic Scene Perception

Authors

Thireesha Suryadevara¹, Mudassir Rafi¹^{, 2}^{, *}, Nandini Mokhamatam¹, Sai Keerthi Aluri¹, Vishnu Priya¹, Harika Kommu¹

¹Department of Computer Science and Engineering, SRM University AP, Amaravati, 522240, India

²Department of Computer Science, King Khalid University, Abha, 62529, Kingdom of Saudi Arabia

^*Corresponding author. Email: mudassir.r@srmap.edu.in

Corresponding Author

Mudassir Rafi

Available Online 18 June 2026.

DOI: 10.2991/978-94-6239-707-1_14 How to use a DOI?
Keywords: Channel Attention Modules; Multi-Scale Feature Fusion Modules; YOLOv8
Abstract: Multi-traffic scene perception is critical for the safe routing and tracking of autonomous vehicles in complex urban environments. This paper proposes an enhanced YOLOv8-based vehicle detection framework for top-view traffic scenes by integrating Channel Attention Modules (CAM) and Multi-Scale Feature Fusion Modules (MSFFM). The proposed model achieves superior performance compared to YOLOv8n (mAP50 = 0.8822), YOLOv7 (mAP50 = 0.8822), and YOLOv9 (mAP50 = 0.8980), attaining an mAP50 of 0.9102 and mAP50–95 of 0.6968. The model is trained and validated on a dataset of 5,766 top-view images containing 2,254 annotated vehicle instances across eight classes. The MSFFM improves multi-scale object detection capability, while CAM enhances feature discriminability. The system demonstrates real-time suitability with an inference time of 4.5 ms per image and a compact model size of 11.47 MB. Although performance limitations are observed for densely packed vehicle classes, the overall results confirm the robustness and effectiveness of the proposed framework for intelligent transportation systems within IoT and AI-driven applications.
Copyright: © 2026 The Author(s)
Open Access: Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

<Previous Article In Volume

Next Article In Volume>

Volume Title: Proceedings of the International Conference on Recent Advances in Intelligent and Sustainable Technologies (RAIST 2026)
Series: Atlantis Highlights in Intelligent Systems
Publication Date: 18 June 2026
ISBN: 978-94-6239-707-1
ISSN: 2589-4919
DOI: 10.2991/978-94-6239-707-1_14 How to use a DOI?
Open Access: Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Cite this article

ris enw bib

TY  - CONF
AU  - Thireesha Suryadevara
AU  - Mudassir Rafi
AU  - Nandini Mokhamatam
AU  - Sai Keerthi Aluri
AU  - Vishnu Priya
AU  - Harika Kommu
PY  - 2026
DA  - 2026/06/18
TI  - Multi-Traffic Scene Perception
BT  - Proceedings of the International Conference on Recent Advances in Intelligent and Sustainable Technologies (RAIST 2026)
PB  - Atlantis Press
SP  - 154
EP  - 166
SN  - 2589-4919
UR  - https://doi.org/10.2991/978-94-6239-707-1_14
DO  - 10.2991/978-94-6239-707-1_14
ID  - Suryadevara2026
ER  -

download .riscopy to clipboard