NutritionVerse-3D

NutritionVerse-3D是一个大规模、高质量、高分辨率的数据集，包含105个3D食品模型，由滑铁卢大学视觉与图像处理实验室和加拿大国家研究委员会共同创建。数据集不仅提供了食品的重量、名称和营养值，还支持通过视图合成生成无限数量的2D图像，适用于任何视角和相机角度。创建过程中，研究团队特别注重速度和一致性，使用iPhone和Polycam应用收集高质量的3D模型。该数据集主要应用于老年人的营养摄入追踪，旨在解决老年人营养不良问题，提高生活质量。

NutritionVerse-3D is a large-scale, high-quality, high-resolution dataset comprising 105 3D food models, jointly created by the Vision and Image Processing Lab at the University of Waterloo and the National Research Council Canada. This dataset not only provides the weight, name and nutritional information of each food item, but also supports generating an unlimited number of 2D images via view synthesis for arbitrary viewpoints and camera angles. During the dataset creation process, the research team prioritized speed and consistency, collecting high-quality 3D models using iPhones and the Polycam application. This dataset is primarily applied to nutritional intake tracking for elderly populations, aiming to address malnutrition among the elderly and improve their quality of life.