Related Experiment Video
Updated: May 27, 2026
07:47
Design and Evaluation of Smart Glasses for Food Intake and Physical Activity Classification
Published on: February 14, 2018
Segmentation Assisted Food Classification for Dietary Assessment.
Fengqing Zhu1, Marc Bosch, Tusarebecca Schap
1School of Electrical and Computer Engineering Purdue University, West Lafayette, Indiana USA.
Summary
This study introduces a mobile app for dietary assessment, using image analysis to identify and quantify food intake. This technology aims to improve the accuracy and ease of tracking diet for better health insights.
Area of Science:
- Nutritional Science
- Computer Vision
- Machine Learning
Background:
- Accurate dietary assessment is crucial for understanding diet-health relationships.
- Current methods can be burdensome and inaccurate.
- Mobile device imaging shows promise for less burdensome dietary assessment.
Purpose of the Study:
- To develop automated methods for food identification and segmentation from mobile device images.
- To improve the accuracy of dietary assessment using computer vision and machine learning.
Main Methods:
- Food images are segmented using Normalized Cuts based on color and intensity.
- Color and texture features are extracted from segmented food regions.
- Support vector machine methods are used for food classification and labeling.
Main Results:
- The developed methods enable automatic food item identification and segmentation from single images.
- Refinement of segmentation based on classifier feedback improves food quantity estimation.
Conclusions:
- Automated image analysis of food intake via mobile devices offers a promising approach for accurate and user-friendly dietary assessment.
- This technology can enhance nutritional research and public health initiatives.
