Convolutional Neural Network for Car Recognition

The code contains a basic car recognizer using keras library.

We have implemented a simple convolutional neural network that achieves over 95% accuracy

In the architecture of the Conv net, we have a following layers:

• Conv layer
• ReLU activation
• Pool layer
• Conv layer
• ReLU activation
• Pool layer
• Fully connected layer
• Softmax layer

Setup:

• Install sklearn using sudo pip install sklearn
• Install keras using sudo pip install keras
• Install skimage using sudo pip install scikit-image

Execute:

python kerasCNN.py