Department of Nutrition and Food Science, College of Agriculture and Natural Resources, University of Maryland, College Park, MD, 20742, USA.
CISPA Helmholtz Center for Information Security, 66123, Saarbrucken, Germany.
Adv Sci (Weinh). 2024 Sep;11(34):e2403578. doi: 10.1002/advs.202403578. Epub 2024 Jul 8.
Addressing the critical need for swift and precise nutritional profiling in healthcare and in food industry, this study pioneers the integration of vision-language models (VLMs) with chemical analysis techniques. A cutting-edge VLM is unveiled, utilizing the expansive UMDFood-90k database, to significantly improve the speed and accuracy of nutrient estimation processes. Demonstrating a macro-AUCROC of 0.921 for lipid quantification, the model exhibits less than 10% variance compared to traditional chemical analyses for over 82% of the analyzed food items. This innovative approach not only accelerates nutritional screening by 36.9% when tested amongst students but also sets a new benchmark in the precision of nutritional data compilation. This research marks a substantial leap forward in food science, employing a blend of advanced computational models and chemical validation to offer a rapid, high-throughput solution for nutritional analysis.
为满足医疗保健和食品行业对快速、准确营养分析的迫切需求，本研究开创性地将视觉-语言模型（VLMs）与化学分析技术相结合。研究中提出了一种先进的 VLMs，利用庞大的 UMDFood-90k 数据库，显著提高了营养估计过程的速度和准确性。该模型在脂质定量方面的宏 AUCROC 达到 0.921，与传统化学分析相比，对于 82%以上的分析食品，其变异小于 10%。该创新方法不仅在学生群体中的营养筛查中提高了 36.9%的速度，而且在营养数据编制的精确性方面也树立了新的标杆。该研究在食品科学领域取得了重大进展，采用先进的计算模型和化学验证的组合，为营养分析提供了快速、高通量的解决方案。
