Real-Time Traffic Sign Recognition Using Deep Learning

Traffic Sign Recognition (TSR) is one of the most sought-after topics in computer vision, mostly due to increasing scope and advancements self-driving cars. In our study, we attempt implement a TSR system that helps driver stay alert during driving by providing information about various traffic signs encountered. We will be looking at working model classifies gives output form an audio message. Our study focused on sign detection recognition Indian roads. A dataset road was created, based upon which deep learning work. The developed deployed NVIDIA Jetson Nano using YOLOv4 architecture, giving accuracy range 54.68–76.55% architecture. YOLOv4-Tiny with DeepStream implementation achieved FPS 32.5, par real-time requirements.