Intelligent Dietary Nutritional Analysis Robot Perception System Based on AI Stereo Vision

In the era of intelligent health management, AI-powered robotics are reshaping the landscape of nutritional monitoring. This study presents HealthMesh, an AI-integrated dietary analysis system designed for mobile devices. Leveraging deep learning and 3D vision, HealthMesh enables users to capture multi-angle food images through smartphones, which are processed in real-time via cloud infrastructure. The system employs YOLOv8 for high-precision food item detection and DUSt3R for dense 3D reconstruction, facilitating accurate volume estimation under real-world conditions. By mapping volumetric and compositional data to a comprehensive nutrition database, the system automatically calculates caloric and nutrient content-including proteins, carbohydrates, fats, vitamins, and minerals. HealthMesh generates real-time nutritional reports and visual summaries, promoting user awareness and decision-making. Through stereo vision and robotic perception algorithms, the platform offers an autonomous and scalable solution for intelligent nutrition tracking, applicable in personal. This work exemplifies the integration of AI-driven robotic perception, real-time processing, and mobile deployment, advancing the capabilities of health-focused automation systems