How do you capture environmental or camera settings in Machine Learning CNN architectures?

Question

I\'ve been looking at computer vision problems. The CNN works exclusively with the image, and other factors like camera settings (f stop, ISO, etc) or environmental factors