Fig. 3
Guidelines for how the dataset can be used in machine learning model training and testing. In order to use the dataset to train a machine learning model, supervised learning can be used, wherein the model will be trained on labeled images. Model performance can then be tested by using new images from this dataset that the model has not yet seen and testing the ability of the model to classify them correctly. Subsequently, the performance of the model can be assessed and further optimized. Once the model is performing as desired, it can be used to classify new clinical images prepared by other researchers and clinicians.
