<![CDATA[Pull-ups are an effective indicator of physical fitness for training upper body muscle strength. However, pull-ups often face problems such as technical errors and inaccuracies in counting the number of repetitions, especially when performed independently without the assistance of a trainer. This study aims to develop and evaluate the effectiveness of a computer vision-based system in automatically detecting, calculating, and validating pull-up movements through video analysis. The system was developed as a website using the MediaPipe framework to detect key points of the body (pose landmarks) and trigonometric methods to calculate the elbow angle as the main parameter in determining the up and down phases of the movement. This study used a quantitative approach with an experimental method, where the system's calculation results were compared with manual calculations for comparison. Test videos of 10–20 seconds with a minimum resolution of 720p were used as system input. The results showed that the system was able to calculate the number of pull-up repetitions accurately and consistently based on changes in elbow angle, and provided objective movement validation. Thus, this system can be a practical, efficient, and easily accessible alternative solution to help users evaluate pull-up exercises independently without the need for additional devices.]]>
Copyrights © 2026
