Train an Age Estimation Network

Combining resources from the Wolfram Neural Net Repository and the surgery capabilities of the Wolfram Language, train a net to estimate the age of a person.

Get an image classification neural net from the repository.

Remove the classification layers and replace them with an age classifier.

This network estimates an age distribution for any given image.

Define a suitable loss function.

Launch a training using an {image -> age..} training and validation set, like UTKFaces.

Estimate the age and plot the age distribution for a face image.