Bite Counter Project


Obesity is a growing problem. In this project, we are investigating a new method to measure eating intake. The Bite Counter device is worn like a watch and tracks wrist motion during eating. By detecting a characteristic pattern, it can identify when a bite of food has been taken. It can monitor intake in real-time and provide feedback to the wearer. The feedback could tell the user to slow down, or tell the user to stop eating after a target intake had been reached, or help the user track long-term eating patterns. The bite counter is a tool to help people track their intake.

Bite Counter device worn on the wrist like a watch Click here to read more


Pattern recognition

When a person takes a bite of food, the wrist of the person rolls. Generally, fingers are aimed downwards to pick something up, and sideways to place it into the mouth. This pattern holds regardless of the type of food or utensil.


Bites as a measure

Isn't a bite of carrots different from a bite of candy? Of course. But nobody gains or loses weight in a single bite, or even a single meal. A common guideline is to lose a pound of weight per week. Our hypothesis is that bite count could serve as a surrogate for calorie count over a period of time. By automating the counting process, and enabling it anywhere, any time, the bite counter can empower individuals to better monitor intake. As with calories, it is possible to count bites for a single meal; but with either measure reduction goals are best evaluated over at least a day.

Bite goals, like calorie goals, should be custom set to the individual. This would be based not only upon the size, gender, age, and activity level of the person, but also based upon the foods typically eaten. For example, a vegetarian may have a higher bite count per day than someone who regularly eats more energy-dense foods. In addition, people tend to eat the same foods week to week. Therefore, setting bite reduction goals over a period of time has the same effect as setting calorie reduction goals.

These graphs show data that relates bite count to calories. Each data point is for one meal. We tracked 83 people for 2 weeks, each graph shows all the meals for a single person. The data on the left show an approximately 0.4 correlation, while the data on the right show an approximately 0.7 correlation. Our study showed that most people correlate bites to calories in this range.

Video

How to use a Bite Counter to measure eating intake for a meal .

A laboratory demo of a few bites, showing a signal-to-bite correlation.

Public relations video for NIH funded study. Thanks to Aramark, SC Launch and CURF for supporting this study.


In the news


Funding:


Papers about this project:


People working on this project:

Group photo, September 2012:

Collaborators at Clemson:

Collaborators at other institutions: Former students who worked on project (theses):

Devices

In development
Version 1 (2007). Tethered sensor. Version 2 (2008). Wireless sensor. Version 3 (2010). Self-contained unit, custom case. Version 4 (2011). Manufactured unit. Version 5 (2014). 2nd generation manufacturing.

Over the years our device has matured from a sensor wired to a desktop computer sitting next to the eater, to a fully self-contained device. Units like the one shown as version 4 are being manufactured at Bite Technologies.


Last updated August 2014

Bite Counter Project Page / Clemson / ahoover@clemson.edu