I have carried out experiments around the techniques of computational photography. Those are used to augment the traditional process of image capture with algorithmic parts. The following example is the result of my own implementation of focal point fusion. You can check out the code and report here.

The following images are used as input to the algorithm:

bookshelf1

bookshelf1

bookshelf2

bookshelf2

bookshelf3

bookshelf3

bookshelf4

bookshelf4

bookshelf5

bookshelf5

bookshelf6

bookshelf6

bookshelf7

bookshelf7

The algorithm detects the most in-focus area of each image, as shown on this map where each color represents an image index:

masks

Masques de fusion

A final fusion step produces the final result:

fusion

Focal point fusion