| Summary: | Traffic signs are a very crucial part in ensuring the safety of people travelling on road and in avoiding accidents. Often, we find various types of signs on the roadsides like speed limit signs, stop signs, yield signs etc. As humans, we can recognize these signs by looking at them, but computer systems cannot interpret what they mean. Traffic sign classification project aims at classifying these traffic signs automatically into their respective types using Convolutional neural networks. This project can then be implemented in autonomous driving vehicles, helping them in making better decisions and providing safer drives. In this project, we use the German Traffic sign dataset which comprises of traffic sign images of forty-three classes. Along with convolutional neural networks, we also use various preprocessing techniques like data augmentation, grayscaling, normalization for enhancing the feature recognition aspect of images.
|
|---|