Development and Validation of a Photographic Food Atlas for Portion Size Estimation and Educational Use in Türkiye

Photographic food atlases are visual aids used to improve portion size estimation in dietary assessment and offer potential use in nutrition education and counseling. This study aimed to develop and validate a photographic atlas for portion size estimation and educational use in Türkiye.

The atlas was developed using standardized procedures for food selection, portion size determination, and image acquisition, and comprised 724 food items represented by 1939 photographs and 2928 portion sizes. Validation was conducted in two phases. In the perception-based phase, atlas-selected and self-estimated portion sizes were compared with presented reference portions, and participants were asked to estimate the weight of the presented portions. In the post-training phase, portion placement accuracy was assessed following atlas-based training. Each phase included 30 participants evaluating 15 food items, resulting in 450 observations per phase and 900 observations in total. Agreement was assessed using Intraclass Correlation Coefficients (ICC) and Bland-Altman analysis, and differences were evaluated using the Wilcoxon signed-rank test.

Atlas-based portion size estimations showed good agreement with presented portion weights (ICC = 0.82, p < 0.001), whereas self-estimated portions demonstrated poor agreement (ICC = 0.49, p < 0.001). Bland-Altman analyses showed a mean difference of -5.5 ± 30.67 g, with limits of agreement wider for larger portions than for smaller portions. Absolute errors were lower for atlas-based estimations compared with self-estimation (p < 0.001). In the training phase, the mean relative portion placement error was 24.0 ± 19.3%, with food-specific values ranging from 15.4 ± 13.8% to 41.4 ± 29.2%.

The developed photographic atlas demonstrated acceptable validity for dietary assessment and potential applicability in nutrition education.