Hi

If you do not mention steps_per_epoch in the training generator, the Imagedatagenerator generates different random augmented images for every batch in each epoch. For example, if you have 500 images and batches=50, in every epoch, the imagedatagenerator generates 10 different augmented image series.

You can also visit our shared code to find useful codes for augmenting data and training:

https://colab.research.google.com/github/mr7495/covid19/blob/master/data_Loading_Training_Evaluating.ipynb#scrollTo=2Gx5rM25d4W4

You can also use save_to_dir in imagedatagenerator.flow function to save and visualize the augmented images in every epoch.

D Vaishnavi

https://stackoverflow.com/questions/51748514/does-imagedatagenerator-add-more-images-to-my-dataset

https://stackoverflow.com/questions/51748514/does-imagedatagenerator-add-more-images-to-my-dataset

https://towardsdatascience.com/image-augmentation-for-deep-learning-histogram-equalization-a71387f609b2

https://towardsdatascience.com/image-augmentation-for-deep-learning-using-keras-and-histogram-equalization-9329f6ae5085