Image-Data-Augmentation-with-Keras

Data augmentation is a strategy that enables to increase the amount of training data using information available from existing training data. To get better generalization in your model you need more data and as much variation possible in the data. Sometimes, dataset is not big enough to capture enough variation, in such cases we need to generate more data from given dataset. That is were Data Augmentation can play a very important role.

Augmentation Operations:

• Rotation
• Width and Height Shifts
• Brightness
• Shear Transformation
• Zoom
• Channel Shift
• Horizontal and Vertical Flips
• Data Normalization
• Rescale and Preprocessing Function
• Using in Model Training