Comparison of digital photography to weighed and visual estimation of portion sizes

The primary goal was to test the validity of digital photography for measuring food portion sizes compared with weighed foods and with direct visual estimation. A total of 60 test meals consisting of 10 different portion sizes from six different university cafeteria menus were prepared and weighed. Food selections and plate waste, as estimated by digital photography and direct visual estimation, were compared with weighed foods. For each method, three observers independently estimated portion sizes of each food. Observers expressed the portion sizes as a percentage of a standard serving. These percentages were multiplied by the weight of the standard portion to yield estimated weights. To test validity, the estimates of food weights derived from both methods were compared with weighed foods using correlations and were compared with each other using Bland-Altman regression analysis. For the digital photography and direct visual estimation methods, estimates of the portion sizes for food selections, plate waste, and food intake were highly correlated with weighed foods. Both methods tended to yield small overestimates or underestimates. Bland-Altman regression found the two estimation methods to yield comparable results (bias less than 1.5 g). These findings support the validity of the digital photography method for measuring portion sizes. Digital photography may be most useful for measuring food intake in settings that allow for the direct observation of food selections and plate waste but require minimum disruption of the eating environment, and allow unhurried estimates of portion sizes.