A research paper co-authored by Prof. Zhang Yingqian and undergraduate student Ouyang Yifan from the School of Artificial Intelligence and Robotics at Xiamen University Malaysia (XMUM) has been published in Signal, Image and Video Processing, a peer-reviewed journal by Springer that focuses on signal processing, image analysis, and video technologies.

The article, titled “ECSNet: A Lightweight and Enhanced YOLOv8-Based Model for Traffic Sign Detection”, introduces a lightweight model designed to address challenges in intelligent transportation systems, particularly real-time video processing on devices with limited resources.

The proposed model, ECSNet, is built on YOLOv8n and integrates two new modules. The Enhanced Connection Path Aggregation Network (ECPANet) improves multi-scale feature fusion while reducing computational complexity, while the Shared Convolution Head (SCHead) compresses the model size without compromising detection performance. Together, these modules enable ECSNet to deliver efficient, real-time traffic sign detection on devices with constrained computational power.

Led by student Ouyang Yifan under the supervision of Prof. Zhang Yingqian, the study demonstrates that ECSNet achieves higher accuracy and efficiency compared to existing methods. The research highlights the potential of lightweight models to improve intelligent transportation systems and embedded applications.

Image