Maintaining a database of food and beverages
One of the biggest challenges with vision based Self Checkout is teaching the system all your food and beverage products. On average, a business restaurant offers upwards of 30 different beverages - cola, energy drinks, juices, water to name but a few.
The checkout system must be able to identify each of these products. To enable identification, the process usually involves uploading images of the product and afterwards annotating them, i.e. telling the system the location of the product in the image.
It can become a very labour intense task and makes the initial introduction (onboarding) of vision based Self Checkout potentially an expensive one. At NUA, we came up with a solution to speed up the process - our Central Food and Beverages database.
Below, we will introduce 2 new drinks to our local outlet database. But before going to the effort of teaching the system the products, we will check the central database to see if one of the other outlets in our organization has already done it. After all, time is money!
Screenshot 1: Introducing 2 new beverages into our local product database, Fritz Apfel-Kirsch Schorle and Granini Apfelsaft.
Currently, both products have no images associated to them. But before we go to the effort of learning them, lets check our central database to see if they are already there.
The workflow is shown in the video below.
Video 1: perform a lookup in the central product database:
Wow! that was easy! Thankfully someone else in the organization had already annotated both products and left us with no work to do.
Therein lies the main advantage of such a central database - each outlet shares the job of maintaining and creating products useful to everyone. And in the specific case of beverages, a significant amount of time can be saved.
There are other uses cases - like storing side dishes that are not specific to any one outlet - pommes frites etc,
Thank you for taking the time to read out article! We hope it was of benefit to you!