YOLO-PPA based Efficient Traffic Sign Detection for Cruise Control in Autonomous Driving

It is very important to detect traffic signs efficiently and accurately in autonomous driving systems. However, the farther the distance, the smaller the traffic signs. Existing object detection algorithms can hardly detect these small-scaled signs. In addition, the performance of embedded devices on vehicles limits the scale of detection modesl. To address these challenges, a YOLO-PPA based traffic sign detection algorithm is proposed in this paper. Firstly, we introduce Parallelized Patch Aware Attention (PPA) into YOLO. PPA utilizes multi-branch feature extraction strategy and spatial-channel attention to capture features at different scales and levels, thereby enhancing the feature extraction capability for long-distance traffic signs. To improve model efficiency without reducing accuracy, we introduce Partial Convolution (PConv) into the YOLO's C2F module. During model training, we introduce APLoss to replace the original classification loss to tackle the serious imbalance of traffic sign categories. The experimental results on the GTSDB dataset show that compared to the original YOLO, the proposed method improves inference efficiency by 11.2%. The mAP@50 is also improved by 93.2%, which demonstrates the effectiveness of the proposed YOLO-PPA.