data Augmentation for resizing images differently to reduce overfitting

Hi All,

thanx for assiting me in advance! i have some code i am understanding and needs some clarification on the pixel range and what it does!

can you help me understand this please!

here's the code!

pixelRange = [-30 30];
imageAugmenter = imageDataAugmenter( ...
    'RandXReflection',true, ...
    'RandXTranslation',pixelRange, ...
    'RandYTranslation',pixelRange);
augimdsTrain = augmentedImageDatastore(inputSize(1:2),imdsTrain, ...
    'DataAugmentation',imageAugmenter);

Answer 1

[Jyothis Gireesh]

“RandXTranslation” and “RandYTranslation” are name-value pairs used to indicate the horizontal and vertical translations to be applied to the images for augmentation. For 2-element numeric vector inputs, the translation value is drawn from a uniform distribution specified by the range given in the name-value pair.

In this case the range is specified by the variable “pixelRange”.

Please refer to the following document link on “imageDataAugmenter” to get a better understanding about the augmentation procedure.

https://www.mathworks.com/help/deeplearning/ref/imagedataaugmenter.html

[Matpar]

Ok, checking the link and thank you for responding to my questions