The automatic recognition of food on images has numerous interesting applications, including nutritional tracking in medical cohorts. The problem has received significant research attention, but an ongoing public benchmark on non-biased (i.e., not scraped from web) data to develop open and reproducible algorithms has been missing. Here, we report on the setup of such a benchmark using publicly available food images sourced through the mobile MyFoodRepo app used in research cohorts. Through four rounds, the benchmark released the MyFoodRepo-273 dataset constituting 24,119 images and a total of 39,325 segmented polygons categorized in 273 different classes. Models were evaluated on private tests sets from the same platform with 5,000 images and 7,865 annotations in the final round. Top-performing models on the 273 food categories reached a mean average precision of 0.568 (round 4) and a mean average recall of 0.885 (round 3), and were deployed in production use of the MyFoodRepo app. We present experimental validation of round 4 results, and discuss implications of the benchmark setup designed to increase the size and diversity of the dataset for future rounds.

The nutritional consequences of progressively replacing meat products with plant-based foods need to be systematically evaluated. Modeling analyses provide insights into the predicted food consumption and nutritional adequacy of plant-based diets. We developed a novel methodology to simulate food patterns and evaluate diet quality. Meal data from the National Health and Nutrition Examination Survey (NHANES) 2017–2018 was used to create 100 7-day meal plans subject to various nutrient and food group optimization criteria. Omnivore (reference diet), flexitarian, pescatarian, and vegetarian food patterns were modeled using mixed integer linear programming. The modeled food patterns used the 25th and 75th percentiles of the US Usual Dietary Intakes to set the optimization constraints. The diet quality was determined using the Healthy Eating Index 2015 (HEI-2015). The modeled vegetarian, pescatarian, and flexitarian food patterns outperformed the omnivore diet on the HEI-2015, with the vegetarian pattern achieving the highest score (82 for females, 78 for males). Modeled flexitarian patterns, with a 25 to 75% reduction in animal protein, offer viable options for those seeking to reduce but not eliminate their animal protein intake while supporting the transition from omnivore to fully plant-based diets. This methodology could be applied to evaluate the nutrient and diet quality of different dietary patterns with various constraints.

No abstract

No abstract