An Accurate Automatic Traffic Signal Detector Using CNN Model

The main objective is to train and build an efficient artificial neural network model, such that the model’s accuracy is high enough to be able to apply it in real life. If the detector is not trained well or rather we can say, if the accuracy of the detector is not high enough, the automated car/driver will end up doing a wrong recognition. This can cause many accidents. Our main aim here is to check and increase the accuracy of the designed model and also, at the same time, ensuring minimal data loss. This can be achieved by proper preprocessing of the data and preventing overfitting. The result shows and compares how the accuracy of the machine learning model increases after proper data preprocessing is performed before feeding it into the model. Also, keeping into account that model overfitting is prevented.