Currently, two main approaches for nutrition tracking are pursued today: (1) Database-driven applications, in which the user looks up the consumed food in a database. While this approach should result in precise nutritional values, entering food is cumbersome for the user e.g. for mixed meals. (2) In photo-driven applications, the user takes a picture of a meal which may be used for reflection and manual analysis, or for image analysis and automatic interpretation. Manual analysis is tedious; image-analysis remains challenging e.g. when differentiating yoghurt from whipped cream. Moreover, taking a picture of every meal is annoying and at least two photos before and after eating is needed.
In this project we would propose a 3rd option; weight-driven nutrition tracking in which you use a scale to measure the intake of food. Jakob Stoustrup – a professor in automation theory at AAU – has argued that just tracking the amount of food consumed is a “good enough” approximation for nutrition tracking. See his blog entry for at description of his “Closed-loop weight control: the power of feedback”.
- Thesis type: System implementation with a small user study.
- Technical skills: Hardware prototyping (e.g. Arduino); System Software design and implementation.
- Research skills: User-centered design; HW/SW co-design; Usability evaluation
In collaboration withCopenhagen Center for Health Technology