I'm new to data augmentation and so far I've gathered that it is used to make the dataset bigger by changing the data in it slightly(i.e. rotating or cropping) and adding the augmented images to the dataset. Does it work that way? And if so, I have seen a lot of examples of using Albumentations library and in those examples we just change images in the dataset with a certain probability, but the size of the dataset remains the same. I feel like I'm missing something. Thank you in advance!
Asked
Active
Viewed 126 times
