Exercise video analysis for health monitoring
Abstract
Our paper explores the development of machine learning techniques for a fitness solution utilizing computer vision and deep learning. The solution addresses two main goals: real-time exercise tracking and also pose classification. The paper discusses the increasing need for real-time fitness applications that actively guide and evaluate users, and the issues individuals face in consistently tracking their health. The developed system uses pose estimation and classification to evaluate exercise form and count repetitions. This paper includes an evaluation of various models for exercise classification using pose estimation.
Fields of Study & Subjects
Human Pose and Action Recognition
Context-Aware Activity Recognition Systems
Hand Gesture Recognition Systems
Publication details
- Journal / source
- IET conference proceedings.
- Volume
- 2025
- Issue
- 43
- Pages
- 211–218
- Publication date
- 1 February 2026
- Language
- EN