The TONI-AI project has been carried out and funded by SMART4ALL Program CTTE Call 1. The TONI-AI consortium consists of 3 partners: 5M ICT, GGZ Zuid-Oost Brabant from The Netherlands and Technical University of Eindhoven from The Netherlands.

Objectives

The project was about AI and healthy living. Currently, nutrition tracking applications are often inaccurate and cumbersome to use. In this project, we proposed an image-based AI solution for nutrition intake calculation, which does three things: it recognizes the food type from the image, calculates where the food is and predicts the quantity of the food. By combining these, our AI solution is able to calculate the nutritional values of the food in the picture

Challenges

The main challenge in this project was to automate calorie intake monitoring by an application which could recognize different food types. We did this by developing and AI solution for food recognition and volume estimation for prepared food, but also a bar code scanner, such that users could easily log packaged foods, such as crackers or juices. The application was developed such that doctors and fitness trainers could access their patients'/clients' dietary habits and offer them advice on these. In order to ensure the privacy of the patients, all AI models are ran locally on a mobile device of the user.

Technology

The innovation in this project was two-fold. We developed an AI solution which can automatically calculate nutrition values from food images. Secondly, we implemented the system in such a way that it ensures privacy to the users. We want to thank the SMART4ALL project for their funding and coaching support which helped us to accelerate the development and create a successful market proposition.

Market

The market targeted by the TONI-AI project consists of healthcare centers for eating disorders and personal fitness trainers. Compared to existing apps, our solution provides an accurate recognition of homemade meals. In addition, our solution also provides involvement of nutrition specialists, who can give the user real-time feedback on their dietary habits. The service will be offered through the FitSprite Nutrition App, which preview is already available on the App Store and the Play Store.

Consortium