
Introduction
Have you ever ever looked for pictures on the computer of a friend? And where you appeared? Then you know how slow and boring it can be, especially with the amount of photographic images that store our computers.
With shareFace now be fast and fun. You just need to install it on two computers, yours and that of your friend. On your machine uses the webcam to make a photograph of your own face, put your computer closer to the computer of your friend, and vía NFC, shareFace shall search in all photos stored in the other computer, retrieving where your face appears.
Technical background
For shareFace do its magic, the development is focusing on the following key components:
-
Windows Storage Search API, to search images on the server.
-
libFace - Face Recognition Library (http://libface.sourceforge.net/file/Home.html/), open source library to face recognition.
-
Windows Networking Proximity API, to transfer information between devices.
Step by step
At first the user must identify in the program as user looking for your images, or as someone who gives your images to another.

If we select the client rol, we will use the webcam to take a picture of our face. (As point of improvement, I will implement a minimum as taking five different photos).

Then we approach the devices, so that the face of the client user is transferred via NFC.

Once received the image of the face, the server device starts searching for photos that match the face, in all those folders where given access permissions.
At the moment that is the first photo found, will start transferring images via NFC, running in parallel search and transfer.

Points of Interest
Every time that the server founds a picture, starts the discovery process of the different faces that appear. Once detected faces, we sliced the photo into as many fragments as found faces, and each of these fragments is subject to verification of the face.
Current state
-
Fragmentation and identification of images, completed.
-
Client UI, and webcam manipulation, done.
-
NFC transfers not implemented yet. Server UI also pending.