If you are on diet, the easiest way to evaluate the progress is to stand up on the scale and keep tracking your weight. However, there are many ways to monitor your dietary. Oliver Amft , Mathias Stager, Gerhard Troster from Switzerland and Paul Lukowicz from Austria proposed a new way to monitor what you eat by analysing chewing sounds. Here is a part from abstract:
We demonstrate that sound from the user’s mouth can be used to detect that he/she is eating. The paper also shows how different kinds of food can be recognized by analyzing chewing sounds. The sounds are acquired with a microphone located inside the ear canal. This is an unobtrusive location widely accepted in other applications (hearing aids, headsets). To validate our method we present experimental results containing 3500 seconds of chewing data from four sub jects on four different food types typically found in a meal. Up to 99% accuracy is achieved on eating recognition and between 80% to 100% on food type classification.
In addition, in their paper, they discussed the three components of imaginary dietary monitoring system, including:
- Monitoring of food intake through appropriate wearable sensors. The main possibilities are
- (a) detecting and analyzing chewing sounds,
- (b) using electrodes mounted on the base of the neck (e.g in a collar) to detect and analyze bolus swallowing,
- (c) using motion sensors on hands to detect food intake related motions.
- Monitoring food preparation/purchase through appropriate environmental augmentation. Here, approaches such as using RFID-tags to recognize food components or communicating with the restaurant computer to get a description and nutrition facts of the order are conceivable.
- Including user habits and high level context detection as additional information sources. Here, one could accentuate the fact that eating habits tend to be associated with locations, times and other activities. Thus information on location (e.g in the dining room sitting at the table), time of day, other activity (unlikely to eat while jogging) etc. provide useful hints.
However, due to technology limitation, they said to The Guardian that
“[t]he system does not need be fully automated to be useful … it is perfectly sufficient if the system can remind the user that, for example, ‘at lunch you had something wet and crisp (could have been salad) and some soft-texture stuff (spaghetti or potatoes)’ and asks him to fill in the details.”