AI-Assisted Nutrition Estimation and Exercise Support for Integrated Lifestyle Management in Patients with Diabetes

Background: /Objectives: Multidimensional lifestyle interventions that combine healthy diet, physical activity, and psychosocial support are central to effective self-management of type 1 and type 2 diabetes. Although digital health tools can improve lifestyle monitoring, existing AI-based dietary assessment methods often struggle to estimate food quantity accurately because food size and scale are difficult to determine from images. This can lead to inconsistent estimates of food mass and energy content. Methods: In this study we introduce nutrition and exercise modules of our AI-assisted Diabetes Care (AIDCare) mHealth digital platform developed for diabetes patients. For effective diet management, we present the PC2FoodNet model, which uses an EfficientNetV2-S backbone with multi-task regression and food-density priors to jointly estimate food volume, mass, and nutrients. For patient safety, we put nutrition-in-the-loop for making diet plans and rectifications of patients’ daily key nutrient intake according to the requirements of the patients. The exercise module includes professionally designed exercises using Unreal Engine’s MetaHuman plugin. The model was evaluated using 4-fold stratified cross-validation on a dataset of 22,070 images covering 40 Turkish food categories. Results: PC2FoodNet achieved a mean Top-1 classification accuracy of 94.39% ± 0.47% and Top-5 accuracy of 98.90% ± 0.18% across the four folds. The corresponding Macro-F1 and Weighted-F1 scores were 93.14% ± 0.59% and 94.16% ± 0.36%, respectively. The physics-constrained regression framework achieved mean RMSE values of 3.01 mL for food volume, 30.17 g for mass, and 57.55 kcal for energy. The confidence-based referral mechanism identified uncertain cases for clinical review, resulting in a clinical referral rate of 23.81%. Conclusions: AIDCare combines physically consistent AI-based dietary assessment with personalized physical activity and continuous psychosocial support. By keeping nutrition involved when AI predictions are uncertain, the proposed framework reduces the risks associated with fully automated lifestyle recommendations. This approach provides a practical foundation for safer, more personalized, and evidence-based diabetes lifestyle management.