Refer to this link to understand the basic working of the project.
- https://www.computervision.zone/courses/traffic-sign-classification/
- https://sid.erda.dk/public/archives/daaeac0d7ce1152aea9b61d9f1e19370/published-archive.html
Note: There have been a number of changes made to the project; it is not an exact replica of the one stated above. In contrast to the reference project, which uses Pickle to store its module files, this project utilises HDF5 (Hierarchical Data Format), a more recent Python package, to preserve its trained modules and integrate it with the most recent version of Python.
A voice announcement function in voicePrompt.py announces the identified traffic song.
pip install numpy
pip install pandas
pip install opencv-python
pip install keras
pip install tesorflow
pip install matplotlib
pip install cv2
pip install h5py
pip install pyttsx3
The project is focused on training and classifying traffic signs using Convolutional Neural Networks (CNNs). The classification is done in real-time using a simple webcam and OpenCV library. CNNs are known for their ability to effectively generalize and classify data with high accuracy, making them suitable for this task.
The project involves training the CNN model using TensorFlow and Keras frameworks. The training data consists of over 35,000 images of 43 different classes of traffic signs. The goal is to build a model that can accurately classify traffic signs based on their visual features.
Overall, the project aims to empower individuals to develop robust traffic sign classification models using CNNs and provides insights into the training process and data requirements for achieving accurate results.