Can you explain how you implemented Convolutional Neural Networks in your fault detection project and the challenges you faced during this process?