SA-YOLOv3: An Efficient and Accurate Object Detector Using Self-Attention Mechanism for Autonomous Driving

• Published in: IEEE transactions on intelligent transportation systems 2022-05, Vol.23 (5), p.4099-4110
• Main Authors: Tian, Daxin, Lin, Chunmian, Zhou, Jianshan, Duan, Xuting, Cao, Yue, Zhao, Dezong, Cao, Dongpu
• Format: Article
• Language: English
• Subjects: attention mechanism; Autonomous driving; Autonomous vehicles; Computer architecture; Convolution; deep learning; Detectors; Feature extraction; Inference; intelligent transportation systems; Object detection; Object recognition; Visualization; YOLOv3
• ISSN: 1524-9050; 1558-0016
• DOI: 10.1109/TITS.2020.3041278

Object detection is becoming increasingly significant for autonomous-driving system. However, poor accuracy or low inference performance limits current object detectors in applying to autonomous driving. In this work, a fast and accurate object detector termed as SA-YOLOv3, is proposed by introducing dilated convolution and self-attention module (SAM) into the architecture of YOLOv3. Furthermore, loss function based on GIoU and focal loss is reconstructed to further optimize detection performance. With an input size of 512\times 512 , our proposed SA-YOLOv3 improves YOLOv3 by 2.58 mAP and 2.63 mAP on KITTI and BDD100K benchmarks, with real-time inference (more than 40 FPS). When compared with other state-of-the-art detectors, it reports better trade-off in terms of detection accuracy and speed, indicating the suitability for autonomous-driving application. To our best knowledge, it is the first method that incorporates YOLOv3 with attention mechanism, and we expect this work would guide for autonomous-driving research in the future.