Data Augmentation To Address Overfitting In Flower Classification CNN In this notebook we will build a CNN to classify flower images. We will also see how our model overfits and how overfitting can be addressed using data augmentation. Data augmentation is a process of generating new training samples from current training dataset using transformations such as zoom, rotations, change in contrast etc
Credits: I used tensorflow offical tutorial: https://www.tensorflow.org/tutorials/images/classification as a reference and made bunch of changes to make it simpler
In below image, 4 new training samples are generated from original sample using different transformations