What is Cheezam ?

Firstly, did you know that there are no less than a thousand different French cheeses? Do you also sometimes find it difficult to differentiate between the different types of "bleu"? You are at a dinner party with friends and you don't know how to recognize a Beaufort from a Comté, and you are afraid to ask which cheese is on the tray? Then use Cheezam! Cheezam is an application that allows you to inform yourself about the cheeses you see! Thanks to artificial intelligence, Cheezam tells you which cheese you just scanned and gives you a lot of information about it!


Who's Cheezam for?


So yes! Whether you are a cheese addict, a lost consumer or a curious tourist for one of the greatest French gastronomic heritage, our Cheezam app will answer all your questions!

via GIPHY


Why download Cheezam?


Cheezam is not just a simple cheese scanner that tells you which cheese it is: a multitude of options are available! Of course, the first feature of Cheezam is to recognize the cheese you are scanning with your smartphone. But here's what you also have access to if you download the application:
changer la liste !! + mettre un exemple de photo In short, Cheezam is the Shazam of cheese.

Convinced ?

via GIPHY


How does it work ?


First of all, how do you usually recognize a cheese?
Before we talk about artificial intelligence, how can we recognize a piece of cheese concretely? Here are a few characteristics that allow us to distinguish according to their type of paste and their characteristics:
  • Soft pasta with a bloomy crust
  • Cooked pressed pasta
  • Uncooked pressed pasta
  • Soft washed-rind pasta
  • Blue-veined pasta
  • Fresh or natural rind cow's milk cheeses
  • Processed cheese makers
  • Goat's cheese
  • Sheep's milk cheeses

An operation based on artificial intelligence...

Cheezam's operation is based on artificial intelligence and, more precisely, on the concept of machine learning. For this application, we mainly use: ml5 and mobile net (see below). In a nutshell: Take a picture and thanks to image recognition by the method of image classification, you get the desired result. We have previously uploaded a database containing the maximum amount of repeated cheese (via several websites) in order to create an efficient pannel. Therefore, once the product is scanned, you get a result that looks approximately like what you can with this example on the reblochon

How does ml5 and mobilenet work more precisely?

Here are some useful explanatory videos to better understand what ml5 and mobilenet are: The following picture explains the type of code we need to write to use ML5.js Webcam Image Classification:

Here, you can better understand what it webcam image classification with ml5


And with this one, you can deeper learn what is image classification with MobileNet


The cherry on the top...

Nous pourrions également penser à une autre fonctionnalité à ajouter à l'application. Cette fonctionnalité utiliserait le concept "databases" qui est le troisième choix dont nous disposions. Rapidement, nous pourrions utiliser cette technologie pour permettre à l'utilisateur de conserver dans un onglet nommé favori ses fromages préférés.
Cet onglet favori constitue en lui même la base de données créée au fur et à mesure par l'utilisateur.
Ce dernier peut: nommer le fromage, mettre des commentaires à son sujet, noter chez quel fromager il l'a acheté ou encore avec quel vin il l'a consommé.
Dès lors, plus besoin de rechercher quel est le nom de ce fromage farfelu mais délicieux qui vous avez goûté, il n'y a plus qu'à vérifier dans vos favoris !
Quickly, we could thanks to glitch and skeleton create a form to link with firebase