Smart Dining Table Project


Concept

Imagine a dining table that could sense foods and beverages being consumed from its surface. The table could calculate statistics about how a person eats including eating rate, bite size, portion consumed and food-to-beverage ratio. All of these habits could be tracked over time to help the user target behavior changes intended to reduce consumption, such as slowing eating, reducing bite size, reducing portion size, or increasing food-to-beverage ratio.


Detecting and measuring bites

The idea of a table that could monitor consumption was first envisioned as something called the universal eating monitor (UEM) in 1980 (see left figure below). It used an embedded scale to measure weight continuously over time as someone ate at the table, calculating g/min consumed. The method works well for detecting individual consumption events (bites) when the user eats under carefully controlled conditions, including not cutting or stirring foods on the table, waiting between bites, and excluding drinks (see middle figure below). We are researching methods that work during natural eating that produces more complex weight signals (see right figure below).

UEM concept (1980) Table weight sensed when eating is carefully controlled Table weight sensed during natural, unrestricted eating


Data

The Clemson Cafeteria data set was used for the experiments described in the below papers.

Papers about this project:


Smart Dining Table Page / Clemson / ahoover@clemson.edu