The augmentation definition is increasing the number of images by using rotation, crop and flip to avoid overfitting. The keras API apply augmentation but no increasing the number of image. What keras augmentation does in images? Is API augmentation such as preprocessing of images? Is augmentation replace the original image with new augmented images?
The Augmentation api apply different augment defined by you (horizontalflip, verticalflip, rotation, zcawhitening etc) to your original image rather than making a copy of it. It apply different augmentation on the same image on different epoch which reduce the generalization error.