Convolutional Neural Networks

Convolution Basics

Convolutional layer

Pooling Layers

Data Augmentation

  • Increasing the training dataset by performing some operations to the input data

  • Ex

    • Shifting images

    • Adding noise

    • Rotating images

    • Flipping images

    • Adding other filters

Transfer Learning

  • Making use of a model that was trained on a very large dataset for a different purpose. The model will likely still have useful training

  • Use the first parts of the nework and retraining the final classification layer