关键词: 交通标志识别；YOLO 深度学习网络；双向特征融合

Abstract: In order to further improve the detection accuracy of traffic sign identification, this paper proposes an improved YOLO deep learning network, which adopts bidirectional feature fusion network, introduces fewer parameters to realize more feature fusion, and calculates target boxes based on Chinese traffic sign data set. Compared with the benchmark network, the results show that the detection performance of the deep learning network is better, which verifies the effectiveness of the improved network.

Key words: Traffic sign recognition; YOLO deep learning network; Bidirectional feature fusion