Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Real-Time Posture Monitoring and Risk Assessment for Manual Lifting Tasks Using MediaPipe and LSTM

Ereena Bagga, Ang Yang

TL;DR

The paper tackles MSD risk in manual lifting by proposing a real-time posture monitoring system that fuses MediaPipe-based pose estimation with an LSTM-based sequential classifier and a risk-assessment module delivered through a web interface. It contributes an LSTM-driven detector for risky lifting postures, a 62-video dataset with MediaPipe keypoints, and an end-to-end webcam-enabled pipeline that provides immediate corrective feedback. The approach achieves high held-out accuracy (categorical: 0.9565; overall: 0.9375) and outperforms domain-specific baselines, demonstrating practical potential for real-time ergonomic interventions and scalable deployment. This work advances MSD prevention by delivering accurate, timely, and actionable posture guidance in real-world lifting tasks, with significant implications for workplace safety and productivity.

Abstract

This research focuses on developing a real-time posture monitoring and risk assessment system for manual lifting tasks using advanced AI and computer vision technologies. Musculoskeletal disorders (MSDs) are a significant concern for workers involved in manual lifting, and traditional methods for posture correction are often inadequate due to delayed feedback and lack of personalized assessment. Our proposed solution integrates AI-driven posture detection, detailed keypoint analysis, risk level determination, and real-time feedback delivered through a user-friendly web interface. The system aims to improve posture, reduce the risk of MSDs, and enhance user engagement. The research involves comprehensive data collection, model training, and iterative development to ensure high accuracy and user satisfaction. The solution's effectiveness is evaluated against existing methodologies, demonstrating significant improvements in real-time feedback and risk assessment. This study contributes to the field by offering a novel approach to posture correction that addresses existing gaps and provides practical, immediate benefits to users.

Real-Time Posture Monitoring and Risk Assessment for Manual Lifting Tasks Using MediaPipe and LSTM

TL;DR

The paper tackles MSD risk in manual lifting by proposing a real-time posture monitoring system that fuses MediaPipe-based pose estimation with an LSTM-based sequential classifier and a risk-assessment module delivered through a web interface. It contributes an LSTM-driven detector for risky lifting postures, a 62-video dataset with MediaPipe keypoints, and an end-to-end webcam-enabled pipeline that provides immediate corrective feedback. The approach achieves high held-out accuracy (categorical: 0.9565; overall: 0.9375) and outperforms domain-specific baselines, demonstrating practical potential for real-time ergonomic interventions and scalable deployment. This work advances MSD prevention by delivering accurate, timely, and actionable posture guidance in real-world lifting tasks, with significant implications for workplace safety and productivity.

Abstract

This research focuses on developing a real-time posture monitoring and risk assessment system for manual lifting tasks using advanced AI and computer vision technologies. Musculoskeletal disorders (MSDs) are a significant concern for workers involved in manual lifting, and traditional methods for posture correction are often inadequate due to delayed feedback and lack of personalized assessment. Our proposed solution integrates AI-driven posture detection, detailed keypoint analysis, risk level determination, and real-time feedback delivered through a user-friendly web interface. The system aims to improve posture, reduce the risk of MSDs, and enhance user engagement. The research involves comprehensive data collection, model training, and iterative development to ensure high accuracy and user satisfaction. The solution's effectiveness is evaluated against existing methodologies, demonstrating significant improvements in real-time feedback and risk assessment. This study contributes to the field by offering a novel approach to posture correction that addresses existing gaps and provides practical, immediate benefits to users.
Paper Structure (15 sections, 6 equations, 6 figures, 3 tables)

This paper contains 15 sections, 6 equations, 6 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Literature Review
  3. Literature Synthesis
  4. Identified Gaps
  5. System Design and Methodology
  6. System Architecture
  7. Dataset Details
  8. Implementation
  9. Results and Comparisons
  10. Visualizations
  11. Comparative Analysis of Posture Detection Models
  12. Existing Work 1: Tennis Shot Identification Using YOLOv7 Pose Estimation and LSTM
  13. Existing Work 2: Yoga LSTM Pose Machines
  14. Comparative Analysis
  15. Conclusion and Future Work

Figures (6)

  • Figure 1: Workflow diagram for a lifting posture state of a worker, depicting the flow of information from the webcam to the AI model.
  • Figure 2: Illustrates the identification of 33 landmarks on the human body through the utilization of Mediapipe.
  • Figure 3: Testing the Mediapipe model for each activity. a Bad Posture, b Good Posture.
  • Figure 4: Performance visualization of the Training and Validation Accuracy over Epochs.
  • Figure 5: Performance visualization of the Training and Validation Accuracy over Epochs.
  • ...and 1 more figures