Data augmentation factor in training a CNN

Question

I training a CNN, many authors have mentioned of randomly cropping images from the center of the original image with a factor of 2048 data augmentation. Can anyone plz elaborate what does it mean?

Answer 1

I believe you are referring to the ImageNet Classification with Deep Convolutional Neural Networks data augmentation scheme. The 2048x aspect of their data augmentation scheme goes as follows:

1. First all images are rescaled down to 256x256
2. Then for each image they take random 224x224 sized crops.
3. For each random 224x224 crop, they additionally augment by taking horizontal reflections of these 224x224 patches.

So my guess as to how they get to the 2048x data augmentation factor:

1. There are 32*32 = 1024 possible 224x224 sized image crops of a 256x256 image. To see this simply observe that 256-224=32, so we have 32 possible horizontal indices and 32 possible vertical indices for our crops.
2. Doing horizontal reflections of each crop doubles the size.
3. 1024 * 2 = 2048.

The center crop aspect of your question stems from the fact that the original images are not all the same size. So what the authors did was they rescaled each rectangular image so that the shortest side was now of size 256, and they they took the center crop from this, thereby rescaling the entire dataset to 256x256. Once they have rescaled all the images to 256x256, they can perform the above (up to)-2048x data augmentation scheme.