AI Fitness Tracker: A Smart System for Pose Detection, BMI Analysis, and Personalized Diet Planning

The increasing interest in health and fitness has resulted in a rise of digital fitness platforms; however, most existing systems provide generic workout and diet recommendations without considering individual differences. This paper presents AI Fitness Tracker, a web-based intelligent platform that integrates computer vision–based pose detection, BMI analysis, and a custom rule-based diet recommendation engine. The system uses MediaPipe for real-time body landmark detection and a custom algorithm for posture evaluation. A rule-based decision system generates user-specific diet plans based on BMI, activity level, and calorie requirements. The backend follows a Spring Boot microservice architecture, secured with Keycloak and deployed using Docker, while ReactJS is used for an interactive frontend. The prototype demonstrates effective posture correction, accurate BMI calculation, and personalized diet suggestions, making the platform a comprehensive and practical fitness assistant. ●