Traffic Sign Classification Using Attention Fused Deep Convolutional Neural Network

Autonomous vehicular technology, also known as self-driving or driverless technology, refers to the innovation that enables vehicles to operate without human intervention. Traffic sign classification (TSC) is a critical component in autonomous vehicular technology, as it allows vehicles to recognize and interpret traffic signs, which is essential for safe and rule-compliant navigation. This work proposes a novel attention-fused deep convolutional neural network (AFDCNN) for TSC. The proposed AFDCNN incorporates the capabilities of ResNet50 and EfficientNetV2 by combining their outputs through a self-attention mechanism which enhances its ability to classify traffic signs. Analysis of the GTSRB, LISA, and MASTIF datasets revealed that the proposed model exhibited superior performance compared to state-of-the-art models, as evidenced by higher scores in recall, precision, F1-score, and accuracy metrics.