Random Undersampling for Imbalancd Dataset

Hello

I have a dataset of multiple classes, however, the dataset is imbalanced.

I know that RUSBoost method is effective for such purposes, however, my dataset includes images, and I want to use them for a CNN.

Could you please help me on how to perform Random Undersampling (RUS) prior to inputing the dataset to the CNN?

Thank you.

[Kenta]

This is an example of over-sampling for imbalanced data. This might increase the accuracy of your task.

https://jp.mathworks.com/matlabcentral/fileexchange/78020-oversampling-for-deep-learning-classification-example