COMPREHENSIVE AUTOMATED MONITORING OF DIETARY INTAKE USING AN INTELLIGENT SYSTEM

Healthy dietary intake is essential for living a high quality lifestyle and has the potential to directly prevent eating-related diseases affecting millions of people worldwide. What and how much we eat and drink has an important impact on our health. Diet-related diseases are a major public health concern and continue to endanger our population’s health and the sustainability of our healthcare systems. Unbalanced dietary intake (relating to both what and how much we eat and drink) can result in metabolic disturbances, malnutrition, overweight, poor mental performance and other medical risk factors like hypertension, and osteoporosis. This can further lead to cardiovascular diseases, type 2 diabetes, hip fractures, liver pathologies, Alzheimer’s disease and cancer. It is evident that these dietary-related issues are not limited to a specific population, thereby having a significant impact worldwide. Hence, this project proposes a comprehensive solution for dietary intake monitoring. As current dietary intake monitoring methods are often labour-intensive and cumbersome by requiring manual input, automated tools will ease this process. At KU Leuven, a smart plate was developed for this purpose. However, since this mainly covers only one aspect of dietary intake (volume/weight estimation respectively), combinations with other sensors/tools are necessary to classify food type and detect intake too. Different approaches have been suggested for this (such as acoustic, inertial or visual) and even smart devices like cups or cutlery have been developed already. Nevertheless, until now most of the solutions are deployed individually rather than fused together. In the proposed project we will aim to do the latter with consumer and/or research grade sensors/devices. Apart from hardware, this project will also include software, in the form of a mobile application or web-portal. This software is used to (i) automatically monitor a person’s dietary intake with a minimum of required user-input, and (ii) turn it into personalised and actionable feedback using current and new recommendation techniques (e.g. self-determination theory, blended care approach). Personalised suggestions for end-users on how to improve their eating behaviour will be provided based on data that has been collected. In addition, visualisation techniques will support the recommendation process with input and feedback. Finally, mobile device implementation will improve the accessibility of the solutions. Generally, the aim is to develop a complete toolset that could be perceived both as a monitor and personal digital coach to stimulate healthy dietary intake.