Scholarlink

Pose estimation is a technique used for specific purposes to detect, predict, and analyze key points of a human object. This study employed pose estimation techniques to detect routine exercises and analyze the repetition and intensity of workout exercises for squats, push-ups, and sit-ups. This optimized the tasks of the experts to monitor the progress of the exercise routine and served as a basis for controlling the optimal limit of the routine exercise. Therefore, this study developed a model built in Python, OpenCV, MediaPipe, and CNN, with an accuracy score of 99%, which was deployed in a Flask API. The model application was evaluated by 30 enthusiastic participants to assess how the application conformed to the five criteria from ISO/IEC 25010 as an initial investigation towards designing a full-fledged application. The results revealed that all criteria received a “Satisfactory” evaluation, with Functionality scoring 4.24%, Reliability scoring 4.15%, Usability scoring 4.24%, Performance Efficiency scoring 4.17%, and Portability scoring 4.24%. In light of these results, there was a need to consider other relevant components, as well as the criteria for each, criterion to further improve this work.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Pose Estimation Detection and Analysis for Workout Exercises Using Convolutional Neural Network

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • King Miles Edrianne A. Ramos,
  • Kenneth B. Basibas,
  • Carl T. Alejandro

摘要

Pose estimation is a technique used for specific purposes to detect, predict, and analyze key points of a human object. This study employed pose estimation techniques to detect routine exercises and analyze the repetition and intensity of workout exercises for squats, push-ups, and sit-ups. This optimized the tasks of the experts to monitor the progress of the exercise routine and served as a basis for controlling the optimal limit of the routine exercise. Therefore, this study developed a model built in Python, OpenCV, MediaPipe, and CNN, with an accuracy score of 99%, which was deployed in a Flask API. The model application was evaluated by 30 enthusiastic participants to assess how the application conformed to the five criteria from ISO/IEC 25010 as an initial investigation towards designing a full-fledged application. The results revealed that all criteria received a “Satisfactory” evaluation, with Functionality scoring 4.24%, Reliability scoring 4.15%, Usability scoring 4.24%, Performance Efficiency scoring 4.17%, and Portability scoring 4.24%. In light of these results, there was a need to consider other relevant components, as well as the criteria for each, criterion to further improve this work.