Posture Estimation and Virtual Gym Assistant using Media Pipe
Summary
No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.
Summarise with VIRAT Free to try with your Science Catalogs account.Abstract
This research presents the design and implementation of a Virtual Gym Assistant that utilizes Google's Media pipe library for accurate and real-time human posture estimation.The system is capable of recognizing fundamental physical activities such as squatting, standing, and sitting, while also performing detailed movement analysis.To achieve this, the assistant integrates deep learning algorithms with Media pipe's pose estimation module to detect body landmarks, calculate joint angles, and classify exercise movements with high precision.The primary objective of this work is to improve workout performance, minimize posture-related injuries, and increase user engagement through continuous automated feedback.Live body movements captured via a camera are processed by Media pipe, which identifies 33 anatomical landmarks on the body.These landmarks are further analysed to compute joint angles and motion trajectories that evaluate exercise accuracy.The extracted features are then fed into various machine learning models, including Logistic Regression, Support Vector Machine (SVM), Naïve Bayes, Decision Tree, and Artificial Neural Network (ANN).A comparative analysis reveals that ANN provides the highest classification accuracy in detecting posture errors and analysing exercise patterns, while SVM and Decision Tree also yield promising results.In addition to posture classification, the assistant offers exercise repetition counting, performance insights, and real-time corrective feedback through a voice-enabled interface.This not only ensures proper execution of exercises but also makes workouts more interactive and motivating.The system demonstrates the potential to serve as a low-cost, AI-powered fitness trainer, enabling individuals to practice safe and effective workouts from the comfort of their homes.
Fields of Study & Subjects
Human Pose and Action Recognition
Context-Aware Activity Recognition Systems
Hand Gesture Recognition Systems
Publication details
- Journal / source
- International Research Journal of Modernization in Engineering Technology and Science
- Publication date
- 30 March 2026
- Language
- EN