Qarbs: Accurate results for estimating carbohydrates

18 · Jul · 2023

Estimating carbohydrates as supreme discipline for people living with diabetes is still depending on gut feelings and experiences. For self-prepared foods it’s most of the time a cakewalk but it becomes more complicated for unknown food items or at friend’s places. The same amount of pasta looks completely different on a small plate than on a bigger plate. To eliminate this source of error our target is offering a reliable digital solution that provides accurate results for an enhanced insulin dosage.

Several neural networks have been developed, trained and tested to be able providing an app that fulfills the needs of its users. Based on a set of pictures taken from the meal in front of you, a 3D model is developed and with this the app defines the volume of a single food item. A curated food database allows the app to define the amount of carbohydrate contented in the food item.

Realizing such a valuable app needs specialists from different areas such as nutrition, computer vision, design, software engineering and regulatory who work closely together. The dedicated team is keen on providing to people living with diabetes a certified medical device that does the carbohydrate counting for them.

Project team: Myriam Tinner (Product Manager, DCB), Melanie Stoll (Nutrition Specialist, DCB), Egemen Vardar&Tobias Maurer (Projektleiter, comerge AG), Andreas Pedroni (Quality Manager, comerge AG), Manuela Miksa&Rafal Wòjcik (UI/UX Designer, comerge AG)

Funding: DCB

App interface for meal scanning

More recent news

Sign up for our newsletter and stay up to date.

* Pflichtfelder
DSGVO *

DCB Research AG

Freiburgstrasse 3
3010 Bern
Switzerland