There is currently no accurate measurement of dietary intake. All current methodologies of assessing food intake have high inaccuracy rates. Yet accurate assessment of nutritional intake is a prerequisite to define the nutritional status, nutritional needs of a population and to monitor the effectiveness of public health interventions to maintain nutritional health. To this end, it is necessary to develop tools that facilitate accurate assessment of nutritional intake in populations without affecting their normal routines. Existing dietary methods are labor-intensive, expensive, and do not report nutritional intake accurately or social hierarchy of food intake. This has been a major weakness in nutritional science and a major problem for planning health policy.
Efforts have been made to ease the collection of information, online questionnaires, have been proposed for user to record the food intake immediately rather than retrospectively. However, recent studies have found that there is a significant discrepancy between the reported food intake compare to actual consumption. Instead of relying on user inputs, new wearable sensing technologies have been proposed to enable pervasive detection of eating episodes and automatic generation of dietary diary. However, relying solely on sensor signals, the proposed technologies can only extract very limited information on food consumed. To quantify details of diet, many computer vision approaches have been proposed recently using smartphone cameras to capture, recognize and quantify food. Although the computer vision approach can enable automatic food recognition, these methods rely on user input and this still leaves the subjectivity of food intake reporting. In addition, such approaches are not applicable in resource-poor Low and Middle Income Countries (LMIC) settings, where reliable access to power and internet are often not available, and people often lack the literacy skills to provide accurate reports.
To enable accurate measurement of individual food and nutrient intake in low and middle income countries, this project aims to develop a passive capturing system for dietary assessments for both adults and children living in Low or Middle Income Countries (LMICs). The system mainly consists of wearable vision sensors, fixed cameras, and a cloud storage server. The goal of the project is to provide a low-cost and robust system for accurate measurement of individuals’ dietary intake.
The project is made possible through a grant from the Bill & Melinda Gates Foundation (Opportunity ID: OPP1171395).