Build an algorithm using deep learning and computer vision

I need to build an algorithm for a mobile app that:
- contains some haircut models
- upload a profile photo (right, left, front and behind) of a person
- shows how a selected haircut model would look on a person in different point of views
- it must consider the size of the haircut (if a person has a short hair and the haircut model has a long hair, the model isn't compatible)
- it have to consider the skin color, head format, hair edges and hair texture and color
- the mash up should be more realistic as possible
- possibility to upload new haircut models and organize it in different sections
- possibility to create a register for each person to upload the photos

The goal is to show how a person would look with the selected haircut model, not just to paste the haircut model in a person.