In this project, I used Python and TensorFlow to classify traffic signs.
Dataset used: German Traffic Sign Dataset. This dataset has more than 50,000 images of 43 classes. Download the dataset from here.
I was able to reach a +99% validation accuracy, and a 97.6% testing accuracy.
- Load The Data.
- Dataset Summary & Exploration
- turning data into tensors
- Getting images and their labels
- turning images into tensors
- Data Preprocessing.
- Normalization
- resize
- Design a Model Architecture. -CNN Model
- Model Training and Evaluation.
- Testing the Model Using the Test Set.
- Python 3.7
- TensorFlow 2.4.1 (GPU support)