Modified Dataset Classes for Classification Task

[michellepi]

I have created modifications to the dataset classes to allow us to replicate the classification tasks in each datasets corresponding work

I suggest two changes:

1. Choice to split manually (existing method) or according to split provided in the csv
2. Choice to return labels for classification task, brain plane for FetalPlaneDataset, and plane and country for AfricanFetalPlaneDataset

[mxochicale]

On my setup I need to use image.cpu().numpy() to solve TypeError: Cannot interpret 'torch.float32' as a data type. Let's check that we have the right data format.

[mxochicale]

Hi @michellepi

Just to let you know that I re-based michelle_clf branch with latest changes in the main branch. Feel free to commit any changes. Thanks, --Miguel

[mxochicale]

Hi @michellepi

xfetus_classifier.ipynb looks great. Would you please help with the following comments. Then we can proceed to merge it.

1. Add introduction section to your notebook as what we did at the beginning of this notebook: https://github.com/budai4medtech/xfetus/blob/michelle_clf/examples/curating_datasets/curating_african_fetal_us_datasets.ipynb
2. Clean notebook as there are some cells that seems to be flooded the notebook e.g., "Evaluate Brain Planes"
3. Create a new path "classifiers" under "examples" and move your notebook there and rename xfetus_classifier.ipynb as baseline-classifiers.ipynb
4. Add a basic README in examples/classifiers.

Thanks, Miguel