Video-based exercise recognition and calorie burn estimation

Due to modern living conditions and unhealthy eating habits, obesity and inactivity-related health issues have become major problems today. Individuals aiming to maintain a healthy lifestyle or lose weight often engage in various sports activities, but performing these activities correctly and accurately calculating the calories burned can be challenging. Incorrect calorie calculations can complicate weight control and lead to a loss of motivation, metabolic issues, fatigue, and performance decline. Therefore, accurately calculating the calories burned during exercise is crucial. In this study, software was developed to calculate the calories burned based on the type of exercise performed, using the individual's video footage. The software is capable of identifying the exercise type, counting repetitions, and tracking estimated calorie burn from video footage. Problems such as video classification, repetition counting, and calorie calculation, which are typically addressed separately in the literature, have been integrated into a holistic approach in this study. Commonly preferred exercise types such as pull-ups, push-ups, running, cycling, squats, and jumping rope have been considered. Experimental results demonstrated high accuracy.