FIAP - Faculdade de Informática e Administração Paulista

ErgoPose AI - Smart Ergonomic Monitoring

Team members:

• Vinicius Salvador RM567159
• Pedro Gustavo RM568262

Professors:

Tutor

• Sabrina Otoni

Coordinator

• André Godoi

Description

ErgoPose AI is an intelligent real-time ergonomic monitoring system that uses computer vision and artificial intelligence to analyze and correct body posture during prolonged activity in front of the computer.

The project was developed as a solution to the growing epidemic of postural problems related to remote work and prolonged use of digital devices.

Identified Problem

Epidemiological Context

According to data from the World Health Organization (WHO):

• 80% of the world population has already experienced or will experience back pain at some point in their lives¹
• Work-related musculoskeletal disorders affect millions of workers worldwide²
• Remote work increased by 159% between 2005 and 2017³, a trend accelerated by the COVID-19 pandemic

Consequences of Poor Posture

• Chronic pain: neck, shoulders and lower back
• Repetitive strain injuries (RSI)
• Reduced productivity and quality of life
• High costs with medical treatments and time off work

Target Audience

• Workers in home office mode
• Professionals who spend more than 4 hours a day in front of the computer
• Students in remote classes
• Gamers and digital content creators

---

Proposed Solution

Overview

ErgoPose AI is a desktop software that continuously monitors the user's posture through the computer camera, using advanced pose estimation techniques with YOLO Pose models to:

1. Detect body keypoints (17 keypoints in the COCO standard);
2. Analyze angles and positions of the neck, shoulders and spine;
3. Generate real-time alerts when inadequate postures are identified;
4. Record session history for evolutionary tracking;

Characteristics

• Non-invasive: runs in the background without interfering with work;
• Real-time analysis: instant feedback on posture;
• Guaranteed privacy: local processing, no images sent to the cloud;
• Detailed history: statistical data for tracking progress;
• Light and efficient: uses an optimized pre-trained YOLO11n-pose model.

Features

Real-Time Analysis

• Neck detection: Alerts on excessive tilt (>45°)
• Shoulder symmetry: Identifies misalignments (>15% difference)
• Spine posture: Monitors back curvature (<70° or >110°)

Visual Interface

• Overall ergonomics score (0-100);
• Color indicators:
  • Green: Ideal posture (Score >= 80)
  • Yellow: Attention (Score 60-79)
  • Red: Danger (Score < 60)
• Skeleton overlay with detected keypoints;

---

Folder structure

fiap_cane_harvest_monitor/
│
├── data/              # Data generated by the application;
│   ├── history/       # Posture history across sessions;
│   └── screenshots/   # Captured screenshots;
│
├── models/ # Downloaded pre-trained models
│
├── src/ # Source code
│   ├── controllers/
│   │   ├── __init__.py
│   │   ├── ergonomic_analyzer.py # Ergonomic analyzer
│   │   └── pose_detector.py      # Pose detector (YOLO Pose 11)
│   │
│   └── camera_app.py # Main application
│
│
├── .python-version
├── .gitignore
├── poetry.lock    # Poetry dependencies
├── pyproject.toml # Poetry project
└── README.md      # Main documentation (this file)

How to run the code

Prerequisites

• Python 3.11 or higher installed (version used: 3.11.9)
• Poetry (Python package manager)
• Git (to clone the repository)

Installation

Step 1: Clone the repository

git clone https://github.com/seu-usuario/fiap-global-solution.git

cd fiap-global-solution

Step 2 (optional): Configure Poetry to create the virtual environment inside the project

poetry config virtualenvs.in-project true

Step 3: Install the dependencies

poetry install

Step 4: Run the application

poetry run python src/camera_app.py

Troubleshooting Camera error: Run the IDE/terminal with administrator privileges Model not found: Make sure you have an internet connection on the first run

Release history

• 1.0.0 - 11/19/2025
  • Initial release
  • Pose detection with YOLO11n-pose
  • Data persistence in CSV

References

1 - World Health Organization (WHO). Low back pain. Available at: https://www.who.int/news-room/fact-sheets/detail/low-back-pain
2 - Canadian Centre for Occupational Health And Safety. Work-related Musculoskeletal Disorders (WMSDs). Available at: https://www.ccohs.ca/oshanswers/diseases/rmirsi.html
3 - FlexJobs. 2017 State of Telecommuting in the U.S. Employee Workforce. Available at: https://www.flexjobs.com/blog/post/flexjobs-gwa-report-remote-growth/
4 - Ultralytics YOLO. YOLOv11 Documentation. Available at: https://docs.ultralytics.com/models/yolo11/
5 - COCO Dataset. Keypoint Detection. Available at: https://cocodataset.org/#keypoints-2020

License

FIAP GIT MODEL by Fiap is licensed under Attribution 4.0 International.