I am a computer science student currently working on my final project, which involves finding a classification-based solution for predicting the head size of a fetus during its third month. Here is a reference for a regression solution utilizing image segmentation with the U-Net model, as illustrated in the Fetal-Head-Circumference-Prediction project.
Unlike that solution, I aim to use classification to detect if the head is of proper size.
My question is what approach is best to split a data set of gray images of fetal heads to classify if one is in abnormal size or not?
Here is a sample database for reference.
My initial thought was to organize the images into age groups, similar to the list at the end of the FetalCns_Final.ipynb file provided in the first link. However, I am uncertain if this is the most suitable approach when dealing with similar images to classify age.