GM Traffic Sign Recognition (General Motors Traffic Sign Recognition System) is an advanced driver assistance feature based on computer vision technology. The system adopts hybrid colour segmentation and customized feature extraction technologies, optimized for the complex environment of Malaysian standard and non-standard road signs (such as diamond-shaped warning signs and circular prohibition signs). The measured recognition accuracy reaches 99.9%, with a response time of only 0.3 seconds, which can reduce the risk of violations caused by driver negligence. Current technology trends show that deep learning models (such as YOLOv8 and RT-DETR) further increase the accuracy to 99.8%, and perform particularly well when dealing with challenges such as lighting changes or road construction signs. The system integrates localized dataset training to ensure accurate identification of Malaysian characteristic road signs such as green highway indicator signs and maroon tourism signs, providing dual safety guarantees for autonomous driving and manual driving.
