MONAI & Generative AI

[agokli]

I have existing project with 3D DICOM images and have been using MONAI 's Compose Transform for data augmentation. I have used UNERT and DENSENET121 based models for training and validations. The data augmentation has improved Validation AUC-ROC score (average of 0.57). But, I need to improve the score and thinking about using MONAI's Generative AI implementation. Therefore, I will like to use Compose Transform as well as Generative AI to have more data for training.

(a) Is this appropriate approach (e.g. using Compose Transform and Generative AI)?
(b) If it is appropriate, what can be pipeline for implementation?
(c) Any tutorial available for such pipeline?

Thanks

I

[UTKI20]

(a) Is this an appropriate approach (using Compose Transform and Generative AI)?

Using both Compose Transform for data augmentation and Generative AI techniques can be a powerful approach to enhance your model's training data and potentially improve your validation metrics. Compose Transform allows you to apply a series of transformations to your existing data, augmenting it to increase variability and robustness. Generative AI techniques, on the other hand, can help generate synthetic data that complements your existing dataset, providing additional samples for training.