Prof.H.O.D. LongeDOSUNMU, Moyinoluwa Ayodele2026-08-272026-08-272020-12-10https://ir.bellsuniversity.edu.ng/handle/123456789/497ix,76pagrs illustration. HardbackTraffic signs are a useful invention to help keep the road safe. Although they were designed for quick and easy understanding by humans, they are not so easily recognizable by machines. A traffic sign detection recognition(TSDR) system is a technology by which a vehicle is able to recognize the traffic signs put on the road. An efficient TSDR system in which its recognition and detection mechanisms was developed based on a convolutional neural network model trained by a dataset populated with German traffic signs. The German traffic signs recognition benchmark(GTSRB) Dataset was used because it contains traffic signs that are similar to the convention adopted on Nigerian roads. five different pretrained convolutional neural networks were trained to build five models for traffic sign recognition, the model's performance was evaluated in terms of the speed at which they were trained and the accuracy at which they were able to classify the traffic signs into 43 different classes. The pretrained CNNs used are AlexNet, GoogLeNet, ResNet, VGG-16 and VGG-19 GoogLeNet performed the best with an accuracy of 95%, ResNet and VGG-19 also achieved a similar level of accuracy but they took longer period to be trained. This research shows that the accuracy of the pretrained networks at identifying traffic signs increases as the number of layers increases, also it establishes that Google Net and ResNet are the most ideal of the five pretrained CNNs experimented on, for traffic sign recognition using the GTSRB datasetenAttribution 4.0 InternationalTraffic sign detectionDissertation