How does face detection/recognition work geometrically?

I have to do this huge paper for my math explorations cl explaining how face detection and recognition works using geometric terms.. any advice would be helpful, I've tried researching but I keep coming across very round about answers.