Evaluating food portion estimation accuracy with multi-angle photographs.

Area of Science:

• Nutrition Science
• Dietary Assessment Methods
• Image Analysis in Food Studies

Background:

• Accurate dietary intake surveys are crucial for nutritional research and public health.
• Traditional methods often rely on self-reporting, which can be prone to inaccuracies in portion size estimation.
• Developing objective tools to improve food quantity assessment is an ongoing challenge.

Purpose of the Study:

• To assess the validity of estimating food quantities using photographs taken from various angles.
• To determine the optimal photographic angles for accurately estimating different food types.
• To enhance the precision of dietary intake surveys through improved visual estimation techniques.

Main Methods:

• Eighty-two adults observed simulated meals comprising six food types (rice, soup, fish, vegetables, kimchi, beverages).
• Participants selected matching food photographs taken from multiple angles (0°, 45°, 70° for solids; 45°, 60°, 70° for beverages).
• Accuracy, underestimation, and overestimation rates were calculated for each food and angle combination.

Main Results:

• Cooked rice showed highest accuracy at 45° (74.4%), improving to 85.4% with combined angles.
• Vegetables' accuracy increased to 53.7% when multiple angles were used (P < 0.05).
• Beverages were most accurately estimated at 70° (73.2%), while soup had lower accuracy and higher overestimation.

Conclusions:

• Food quantity estimation accuracy is dependent on food type and photographic angle.
• A 45° angle is optimal for solid foods, while 70° is best for beverages.
• Combining multiple photographic angles significantly improves food quantity estimation accuracy for most food types.