AI Digitized 3D-Printed Meal for Personalized Nutrition

Personalized nutrition requires practical methods to convert dietary assessments into tailored food portions. Current image-based assessment tools and 3D food printing systems are often disconnected, lacking end-to-end evaluation. This study presents an integrated platform that converts meal images into personalized 3D-printed macronutrient supplements. The system combines smartphone image analysis, volume calculation, nutrient recommendation, and automated G-code generation for a custom dual-nozzle 3D printer. Four nutrient-rich inks (avocado, purple sweet potato, red lentil, chicken breast) were formulated with guar gum and characterized for extrusion-based fabrication. The AI dietary assessment module achieved a mean volume and weight estimation error of 37.25%, with macronutrient estimation errors between 31% and 38%. Ablation studies demonstrated that user-calibrated 3D depth reconstruction reduced volume errors compared to 2D baselines and uncalibrated models. During fabrication, the optimized inks exhibited shear-thinning behavior and elastic-dominant responses, enabling stable extrusion. Printed shape fidelity ranged from 89.67% to 99.00%, and weight accuracy from 79.00% to 96.00%. These results demonstrate the feasibility of a closed-loop workflow linking AI dietary assessment with automated, on-demand food fabrication, paving the way for precision nutrition interventions.