Regressing images to a float

Notebook here, open with collaboratory

Is it possible to regress images to a float? For example, images of people regress to their height? Or a person’s face to their age?

I have a number of images of eyes with the corresponding “Pupil-Limbus-Ratio” (PLR) values. This is a value which can be used to tell how dilated a person’s pupil is.


In the model I have run, I’m not sure how to interpret the predictor values. Nor am I sure that the model has been set up correctly so that it does indeed regress these images to their floats.

Can someone with experience weigh in on this?

Hey, I just went over your notebook. I don’t see any problems with the setup as such. I saw just one error which was when you tried to plot confusion matrix. Confusion matrices are for classification problems not regression problems :slight_smile: Rest of it seemed fine. Check with the lesson 3 image regression notebook if you have any doubts.