An Application-Based Ai Fitness Trainer Integrating Multilayer Perceptron Network & Computer Vision

As sedentary lifestyles increase, the occurrence of diseases associated with physical inactivity raises for effective and affordable fitness interventions. Modern fitness applications still lack to provide a range of solutions with real-time pose estimation, personalized recommendations, and confidential management of user subscriptions. This paper proposes a Data-Driven AI Fitness Trainer that integrates real-time pose estimation, hybrid recommendation systems, and blockchainbased subscription models to tackle such gaps. The system combines AI frameworks such as MediaPipe and CVZone for real-time posture analysis. Multilayer perceptron (ML‘P) models are incorporated to provide personalized workout and diet recommendations. A blockchain-based subscription will permit safe, transparent, and efficient transaction management. The proposed system provides a smooth fitness experience by giving real-time feedback on posture correction, personalized fitness regimes, and secure access management.