Using One-Shot Prompting of Non-Fine-Tuned Commercial Artificial Intelligence to Assess Nutrients from Photographs of Japanese Meals
The simplest use of current AI via one-shot prompting and no fine-tuning accurately assesses fiber content in meals but is inaccurate for other nutritional parameters.
Authors
- Toshimasa Yamauchi (ORCID: https://orcid.org/0000-0003-4827-6404)
- Masaomi Nangaku (ORCID: https://orcid.org/0000-0001-7401-2934)
- Kazuhiko Ohe (ORCID: https://orcid.org/0000-0002-4296-9536)
- Y. Y. Ji (ORCID: https://orcid.org/0009-0003-2123-6311)
- Kayo Waki (ORCID: https://orcid.org/0000-0003-0046-2523)
Institutions
- The University of Tokyo (JP)
Publication Details
- Journal
- Journal of Diabetes Science and Technology
- Published
- 2025-01-10
- DOI
- https://doi.org/10.1177/19322968241309889
- Citations
- 1
- Primary Topic
- Nutritional Studies and Diet
- Type
- article
- Field-Weighted Citation Impact
- 1.88