Objective: In diet surveys, quantitative underestimation of food consumption may be due to intentional misreporting or false portion-size reporting. Perception of food photographs used as aids for assessing the actual amounts may have an effect. This study was carried out to assess the validity of food photographs. Design: A real-time test protocol where 52 presented food servings were compared against photographed portions with similar food items. Subjects: Volunteers from the Rehabilitation Company Petrea (in Turku) were recruited, 161 adults participated, and for 146 subjects, complete data were collected. Methods: The proportions of correct estimations and reporting errors, in weights and percentages, are presented by gender and food group. Food descriptors, portion-size options and subject characteristics were studied as potential determinants of accuracy in portion-size estimation. Results: The total proportion of exactly correct estimations was 51% in men and 49% in women. The overall reporting error was À10 g in men and þ 1 g in women for the 52 food servings. Underreporting was typical for bread, spread and cold cuts and dishes in both genders. Over-reporting was typical for cereals in both genders and for snacks, vegetables and fruit in women. The estimation error was associated with the portion-size options but not associated with the energy density of food items, education or body mass index. Conclusions: Food portions in photographs seem to be a useful aid for the quantification of most food items. However, validation studies are needed to test the applicability of photographs for estimating current portions and for searching better tools in dietary surveys.