Smart Food Calorie Prediction: Hybrid CNN-ML Model for Accurate Meal Tracking

The food calorie prediction systems depend on image recognition, basic Machine Learning (ML) models, and manual input without adequate portion size analysis. It is well-recognized that these algorithms frequently identify food categories, portion sizes, and calorie counts incorrectly. The study introduces a unique hybrid model for automated meal tracking that uses Convolutional Neural Networks (CNN), machine learning, and visual feature extraction to achieve improved accuracy in calorie prediction, portion size estimation, and food recognition. It calculates the real calories by combining them with a nutritional database and using depth estimates to establish portion sizes. The results show that the proposed system outperforms the existing system in terms of accuracy (92.4%), precision (91.3%), and recall (93.2%) while requiring less time (1.10s). These demonstrate that a real-world system may be implemented to provide a scalable and well-recognized tool for diet tracking and health management