| Paper Abstract and Keywords |
| Presentation |
2025-06-18 16:40
Development of a Nutrition Estimation Model via Ingredient Estimation from Meal Images for Dietary Counseling by Dietitians Ryoma Maeda, Kiyoharu Aizawa (UTokyo), Keisuke Shirai, Hirotaka Kameko, Shinsuke Mori (Kyoto Univ.), Akiho Shinagawa (UT IHSS ARIHHP), Keiko Namma-Motonaga, Akiko Kamei (JISS), Yoko Yamakata (UTokyo) |
| Abstract |
(in Japanese) |
(See Japanese page) |
| (in English) |
Nutritional support for athletes requires precise estimation of not only major nutrients but also micronutrients. In this study, we propose a method to estimate ingredients and their weights using a large multimodal model, and then to estimate their nutritional values by comparing them with the Standard Tables of Food Composition in Japan. We will attempt to construct the above method by fine-tuning a large multimodal model using annotated data for recipes collected from the Web, and evaluate its performance using recipes with meal images and nutritional data created at a cafeteria for athletes at the Japan Institute of Sports Sciences. |
| Keyword |
(in Japanese) |
(See Japanese page) |
| (in English) |
Multimodal Large Language Model / nutrition estimation / ingredient estimation / / / / / |
| Reference Info. |
ITE Tech. Rep., vol. 49, no. 17, HI2025-19, pp. 1-6, June 2025. |
| Paper # |
HI2025-19 |
| Date of Issue |
2025-06-11 (HI) |
| ISSN |
Online edition: ISSN 2424-1970 |
| Download PDF |
|