Oliver Amft

Home  Projects Publications Events Contact

Activity and Context Embedded Systems Software Tools
Partnerships

Automatic Dietary Monitoring (ADM)

Goals Sensing & Recognition Chewing Analysis Events & Links

Recognising food intake and dietary behaviour without manual logging. (2004-present)

Relevance
A balanced diet is a vital aspect of human health. The prevalence of various forms of malnutrition, including obesity, bulimia, and anorexia, clearly testifies that much is at stake for our health. It is commonly agreed, that a healthy and preventive lifestyle as well as early diagnosis could systematically fight the origin of a variety of chronic diseases, including those related to diet - however achieving and maintaining such a lifestyle is a life-long challenge. We expect that technology can help!

Long-term continuous behavioural monitoring with personalised feedback and coaching from health professionals, dietitians as well as community building, will contribute significantly in this effort. In the Automatic Dietary Monitoring (ADM) project we develop solutions for automatic food intake detection and diet-related behavioural analysis without requiring the individual to log all diet- and food-related details as it is commonly done when using food intake questionnaires. Such questionnaires are required for weight management programs.


Objectives


Sensing and Recognition Approach
Nutrition is a complex behavioural process, composed from diverse activities, blended with physiologic and psychological responses. No sensing technology is capable of acquiring the entire nutrition process and its structure. Hence, the decomposition into relevant sub-problems is required. Recognition methods are be evaluated for a broad set of individual phenomena (I call them "sensing domains"). The solution...

My PhD thesis on the topic is available here Support independent publishing: buy this book on Lulu.




Funding acknowledgements

I am grateful for the funding support provided by:





last update: 2009-02-13