I would like to develop a program that will output past and future face images, if a face image is supplied as input. That is, when I supply an image, the system should:

1. Recognize the face in it.
2. Calculate the approximate age from the input face image.
3. Output two face images, one with 10 year (just saying) younger from the current age and the other 10 year older.

To train my machine learning program I need a face dataset having multiple face images of people sorted according to age.