Design And Implementation Of Android Fitness Application Based On Gesture Recognization

The rise of sports and fitness apps and We-media has given people more fitness options.More and more people choose to exercise at home or in the gym.Compared with expensive private classes,this kind of fitness is more economical and popular with many young people.The following problems are how to ensure that the movements are in place to play a real fitness role when exercising on their own,and how to avoid muscle strain or excessive strain caused by irregular movements when exercising.In response to these problems,with the purpose of improving users’ fitness literacy,a smart fitness application that can monitor and correct users’ fitness movements in a timely manner is designed and studied.It is mainly suitable for mobile devices,such as mobile phones or tablets,to help users develop a good fitness.Fitness habits.The system adopts the Android hybrid development mode,which will reduce the cost of application upgrades.Integrate the Tensorflow Lite gesture recognition framework with the Android application to perform gesture recognition,and use the gesture correction algorithm to identify the user’s current correct gesture to perform gesture correction.This method of gesture recognition and correction not only has high scalability,for example Upgrading gesture recognition and correction models is less costly and does not require additional sensors.After unit testing,integration testing,stress testing and compatibility testing,etc.,the parameters of gesture recognition algorithms for various types of actions are optimized,so that the performance and functions of the system tend to be perfected,and the accuracy of gesture recognition is continuously improved.Compared with the current fitness applications on the market,the system not only realizes the basic functions of user information management,physical information management,sports community,and formulation of exercise plans,but the biggest innovation and advantage lies in the realization of fitness action supervision and Fitness action correction function.The system integrates Tensorflow Lite posture recognition technology,which has realized real-time supervision and correction of the user’s basic fitness movements,and avoided as much as possible the damage to the user’s body caused by incorrect posture during exercise.At the same time,the static posture correction algorithm of the system can support various static fitness movements by adjusting the parameters of different fitness movements,and has good scalability.Dynamic posture correction also achieves a certain amount of wrong action reminders by matching the standard action curve,and also uses the dynamic fitness action counting function to ensure the user’s daily exercise volume.