Making the Most of Mobile Technologies to Estimate Dietary Intake
Principal Investigator: Carol J. Boushey, PhD, MPH, RD
Department of Foods and Nutrition
West Lafayette, IN 47907
[Beginning May 2011]
University of Hawai'i Cancer Center
What's the problem?
To accurately monitor food and nutrient intakes and identify those at risk as
well as those who are meeting recommendations, it is necessary to precisely
estimate intakes. That has consistently posed a challenge because people do not
always report accurately and because dietary assessment instruments contain some
degree of error. In addition, completing and analyzing self-report instruments
can be burdensome for respondents and researchers.
[Return to top]
How has this research addressed the problem?
Dr. Boushey and her colleagues are taking advantage of the near-ubiquitous
use of mobile telephone technology to develop, implement, and evaluate a quick,
convenient, and accurate method for collecting dietary information.
The project involves several components: 1) developing automated image
analysis methods that will identify foods and beverages and estimate the
quantities of these items, 2) modifying the USDA Food and Nutrient Database for
Dietary Studies (FNDDS) to support the system's data, and 3) developing
user-friendly mobile telephone applications that fit the lifestyle of its
Developing Automated Image Analysis Methods
This method, a mobile telephone food record (mpFR), uses a cell phone's
camera to capture images of a food or beverage before and after consumption.
(The "after" image shows whether or not any food or beverage remains after
consumption, and if so, how much.) These images are sent to a central server
for processing. Dr. Boushey's team has developed methods by which the images
can be identified using the food or beverage's unique characteristics, such as
color and texture, and contextual information, such as time of day. The volume
of the food or beverage is estimated by using pre-set geometric templates.
Images of the system are available at www.tadaproject.org.
Modifying FNDDS to Support the System's Data
USDA's FNDDS food composition database does not contain complete density
information, which is needed to translate volume estimates in cubic centimeters
to gram weights. Therefore, Dr. Boushey's team is developing algorithms to
predict densities that can be added to the FNDDS. This will allow investigators
to match the foods and beverages captured by the mpFR system with their
appropriate nutrition information.
Developing Mobile Telephone Applications
Dr. Boushey and her colleagues also have created an easy review application
that allows the user to confirm or change the automated labels provided on the
foods and beverages. The resulting data, electronically provided to
researchers, detail foods and beverages consumed along with energy and nutrient
Testing the System
The investigative team developed a prototype of the mpFR that runs on iPhone
3Gs, and tested it in a convenience sample of 57 men and women, ages 21 to 65
years. This pilot was designed to see how well the technology works and to help
quantify the error associated with the food and nutrient output. Study
participants used the mpFR during a controlled meal session, and 24 (42%)
returned for a second meal session. Participants were able to capture usable
images for analysis and responded favorably to the process of collecting these
data. They also provided useful feedback for improving the accuracy and ease of
use of the mpFR.
In a second study, 12 free-living men and women, ages 20 to 58 years, used
the mpFR for 3 non-consecutive days to capture reported energy intake. The
difference between the mean reported energy intake and mean estimated energy
requirements was -337 kcal (p=0.165), or 87% of estimated energy requirements.
These results reinforce the use of image-based dietary assessment in clinical
and research settings and its potential for leading to more accurate accounts of
[Return to top]
Significance of the research & results
This study is one component of NIH's Genes,
Environment, and Health Initiative (GEI). GEI, begun in 2007, supports the
development of new tools to determine the genetic and environmental roots of
common diseases. Dr. Boushey's research is one of the studies within the GEI's
Measures of Diet and Physical Activity Program Area, which the National
Cancer Institute co-leads with the National Heart, Lung, and Blood Institute.
Results will contribute significantly to current efforts to refine and improve
the assessment of usual dietary intake, both by improving the accuracy of
self-reported dietary data and by reducing the burden on respondents and
researchers of data collection and analysis.
Moreover, this project represents a unique and highly productive
collaboration by a range of scientific disciplines. To address the complexities
of this project, Dr. Boushey has assembled an investigative team that includes
experts in electrical engineering, information science, nutritional
epidemiology, food composition databases, stable isotopes, and statistics.
[Return to top]
Recent related publications of interest
Boushey CJ, Kerr DA, Wright J, Lutes KD, Ebert DS, Delp EJ.
Use of technology in children's dietary assessment.
Eur J Clin Nutr 2009 Feb;63 Suppl 1:S50-7.
Six BL, Schap TE, Zhu FM, Mariappan A, Bosch M, Delp EJ, Ebert DS, Kerr DA, Boushey CJ.
Evidence-based development of a mobile telephone food record.
J Am Diet Assoc 2010 Jan;110(1):74-9.
Zhu F, Bosch M, Woo I, Kim S, Boushey CJ, Ebert DS, Delp EJ.
The Use of Mobile Devices in Aiding Dietary Assessment and Evaluation.
IEEE J Sel Top Signal Process 2010 Aug;4(4):756-766.
[Return to top]