Overcome medical image data scarcity by data augmentation techniques: A review
Laila El Jiani, Sanaa El Filali, El Habib Benlahmer
Abstract
Deep Learning brings back today valuable answers to many issues relating to different areas of life. To maximize the performance of a deep learning model, the model must be trained on a large amount of data. The scarcity of data is the first barrier to developing powerful medical image analysis-based deep learning systems. To overcome data scarcity, data augmentation techniques can be used to increase the diversity of the training dataset. This review summarizes most data augmentation techniques and exposes their performance before and after applying data augmentation. It also demonstrates the positive impact of using these techniques to overcome the lack of data in the medical imaging field.