Abstract
Traffic sign detection and recognition (TSDR) is a critical aspect of autonomous driving and intelligent
transportation systems. Traditional methods of traffic sign detection rely on handcrafted features and classical machine
learning algorithms, which often struggle to achieve high accuracy in complex real-world environments. In contrast,
deep learning techniques, particularly Convolutional Neural Networks (CNNs), have shown remarkable performance in
both detecting and recognizing traffic signs in diverse conditions. This paper reviews the application of deep learning
methods for TSDR, focusing on recent advancements in the use of CNNs, Region-based CNNs (R-CNNs), and YOLO
(You Only Look Once) for real-time detection and recognition of traffic signs. We discuss the challenges, including the
variability of traffic signs, environmental factors, and computational requirements, and present solutions proposed in
the literature. The paper also highlights future directions for improving the robustness, accuracy, and speed of traffic
sign detection and recognition systems, with an emphasis on real-time applications.