A traffic sign recognition system design for road safety assistance

This article outlines the design of a traffic sign image recognition system for road safety assistance. Initially, we used YOLOv7, a computer vision technology based on neural networks, to train an image recognition model on standard traffic sign images in Taiwan. Next, we developed a mobile app using C# and the Unity software development tool and integrated the trained model. This app turns the mobile phone into a road sign image recognition system, providing essential traffic sign information for drivers and visually impaired pedestrians to enhance traffic safety. The article also presents our progress in image training and the real-world performance of the recognition system.