4

How would you solve the problem of identifying certain customer in a grocery store? Suppose our client is already signed-up on our website with an unique ID given to him. To come in to the store, firstly he must place a phone with a QR code in front of the reader, so that our server is informed on a specific client entering the shop. Then after successful authentication he heads towards the shelf with goods and pick some item. Identifying product name turned out not to be a challenge, as opposed to person.

From my perspective, the solution is to make a several photo of face from a different angle while person coming in. Quickly train CNN and feed it with face images when customer picks goods in order to choose the corresponding one.

So, what are your thoughts on this issue? What approach would you take to work it out?

DukeZhou
  • 6,237
  • 5
  • 25
  • 53
Taras Mykhalchuk
  • 149
  • 2
  • What if she will wear sunglasses? What if she will have some makeup? What if he is not shaved today? What if she returned from holidays and is sunburned? What if he was hit on his eye and have a limo? – kakaz Dec 10 '17 at 19:04
  • The point is to train network in a short term period: this time frame is made up of entry point up to the shelf with goods. – Taras Mykhalchuk Dec 11 '17 at 08:19

1 Answers1

2

You might consider pre-training a CNN on a large dataset. The CNN should be structured such that you input 2 different images and the CNN predicts whether or not they are the same person. Your dataset should include images from multiple angles, with and without occlusions like sunglasses, and with changes in hair. (One dataset useful for this is the AR Face Database.) Then you can just check if the face matches any of the pictures you took as they entered the store.

Default picture
  • 212
  • 1
  • 4
  • 1
    This answer might be improved with some links, but it looks like good, general advice regardless, so thanks for contributing! – DukeZhou Mar 09 '18 at 19:07