Question

Instructions
You are applying a deep learning model to detect whether or not your
size of your dataset is small, so you decide to do data augmentation.
cannot be considered as data augmentation?
Select the single best answer:
A. flipping the image either horizontally or vertically
B. changing the contrast, hue, or saturation
C. applying shear to the image
D. cropping the image
E. normalizing images by subtracting the mean image

Answer 1

Final answer:

Data augmentation techniques enhance model performance, with normalizing images being a preprocessing step.

Explanation:

Data augmentation plays a crucial role in improving model performance by increasing the diversity of the dataset through various techniques.

Among the given options, normalizing images by subtracting the mean image cannot be considered data augmentation as it is a preprocessing step to standardize the data.

On the other hand, flipping the image, changing contrast, hue, or saturation, applying shear, and cropping the image are classic examples of data augmentation that help in training deep learning models effectively.

Learn more about Data augmentation techniques in deep learning here:

https://brainly.com/question/44266221