Related Experiment Video
Updated: Jul 15, 2025
13:19
Deep Neural Networks for Image-Based Dietary Assessment
Published on: March 13, 2021
9.2K
DeepTrayMeal: Automatic dietary assessment for Chinese tray meals based on deep learning
Jialin Shi1, Qi Han1, Zhongxiang Cao1
1School of Computer and Communication Engineering, University of Science and Technology Beijing, Beijing, China.
Food Chemistry
|September 24, 2023
Summary
This study introduces ChinaLunchTray-99, the first Chinese tray meal dataset, enabling automatic dietary assessment. The developed framework accurately identifies dishes, estimates volume, and maps nutrition for public health applications.
Area of Science:
- Computer Vision
- Nutrition Science
- Public Health
Background:
- Tray meals are a prevalent dietary pattern in China, necessitating automated methods for dietary assessment.
- Existing research lacks comprehensive Chinese tray meal datasets and effective assessment methodologies.
Purpose of the Study:
- To establish the first large-scale Chinese tray meal dataset (ChinaLunchTray-99).
- To develop and validate a novel framework for automatic dietary assessment of Chinese tray meals.
Main Methods:
- Collected 1185 real-world tray meal images, annotated with bounding boxes and categories for 99 dishes.
- Developed a tray meal detection model achieving 92.13% mean Average Precision.
- Proposed an automated system for volume estimation and subsequent nutrition mapping.
Main Results:
- The ChinaLunchTray-99 dataset provides a valuable resource for research.
- The developed detection model demonstrates high accuracy in identifying dishes within tray meals.
- The integrated framework enables accurate volume estimation and nutrition mapping from images.
Conclusions:
- This work presents a significant advancement in automatic dietary assessment for Chinese cuisine.
- The established dataset and methodology can drive public health initiatives and nutritional research.
- The framework has the potential to be widely adopted for dietary analysis.
