Traffic Sign Classification using Convolutional Neural Networks (CNN)

This script implements a CNN model to classify traffic signs using the GTSRB dataset. The steps include loading and preprocessing the data, defining and training the model, and finally evaluating its accuracy.

Overview

The script performs the following tasks:

1. Import Libraries: Imports necessary libraries for data processing, model creation, and evaluation.

2. Data Loading and Preprocessing:

   • Loads training images from the dataset.
   • Converts images to arrays and labels them according to their respective classes.
   • Splits the dataset into training and testing sets.

3. CNN Model Definition:

   • Defines a Sequential CNN model with multiple convolutional, pooling, and dropout layers to prevent overfitting.
   • Compiles the model using the Adam optimizer and categorical crossentropy as the loss function.

4. Model Training:

   • Trains the model on the preprocessed data with a specified batch size and number of epochs.
   • Uses a validation split to monitor the model's performance during training.

5. Model Evaluation:

   • Loads test data from a CSV file.
   • Preprocesses the test images and predicts their classes using the trained model.
   • Calculates and prints the accuracy of the model on the test data.

Usage

Ensure that the GTSRB dataset is organized correctly with train and test folders containing the images. The test data should also be provided in a Test.csv file with the correct structure.

Run the script to train the CNN model on the training data and evaluate its performance on the test data.

Dependencies

The script requires the following Python libraries:

• os, sys, numpy, pandas
• PIL.Image for image processing
• sklearn for model evaluation and data splitting
• keras for building and training the CNN model

Results

The script will output the accuracy of the model on the test dataset, providing an indicator of its performance in classifying traffic signs.