Image-assisted Dietary Assessment Methods

Accurate dietary assessment is central to the nutrition care process and understanding of underlying diet-disease relationships. Image-assisted food diary in the form of mobile application (app) offers the advantages of convenience and gathering of real-time food intake information. This study describes the evidence-based development of an image-assisted electronic food diary, namely Individual Meal-based Assessment Snapshot (IMBASTM) as a valid dietary assessment tool for Malaysians. 

The prototype was designed iteratively, incorporating user feedback and guided by an agile development model. The app captures multiple-day food records through food images and text and sends this dietary information to a central server for nutrient analysis by researchers. Based on findings of focus group interviews and usability testing, the app features were further enhanced to include customizable reminders, in-app tutorial, and barcode scanning function. IMBASTM was field-tested by triangular comparison with paper-based food diary and repeated 24-hour diet recalls in a cross-over study. The app showed good reliability with validity coefficients ranging from 0.4-0.7. However, estimation accuracy was highly dependent on the reliability of human analysts in assessing dietary intake based on food images.  

IMBASTM is a user-informed image-assisted dietary assessment mobile app that demonstrated the usability and relative validity.  Further work should focus on improving image-assisted dietary assessment skills and establishing the absolute validity of the new tool using biomarkers.