Paper Title
TRAFFNET: A Real-Time Recognition System for Traffic Signs

Nowadays, with so much traffic on the roads, it might be challenging for drivers to read the traffic sign boards while keeping their eyes on the road. The high speed of cars often causes drivers to overlook the traffic sign boards, which can result in breaking the traffic law or causing accidents. Additionally, there are so many traffic signs that drivers could lose track of what each one means. The need for an autonomous system that can recognise different traffic signs and read them aloud to the driver while they are driving is therefore urgent. This will allow the driver to focus solely on the road and not have to worry about the traffic signs. The end-to-end real-time traffic sign categorization system proposed in this research uses a Raspberry Pi 4 connected to a camera to capture traffic signs and recognise them using YOLOv3. After that, a deep learning model is used to recognise the meaning of the traffic sign and translate it into voice using the image of the traffic sign that was just captured. The results demonstrate that MobileNet is the best model for recognising traffic signs with 99.72% accuracy, according to the dataset it was trained on. Keywords - Traffic Sign Recognition, Machine Learning, Deep Learning, Speech Conversion, Transfer Learning