Skip to content

yakupzengin/fitness-trainer-pose-estimation

BranchesTags

Repository files navigation

🏋️ Fitness Trainer with AI Pose Estimation

An AI-powered web application that tracks your exercises using computer vision and provides real-time form feedback with scoring.

Python Flask MediaPipe License

✨ Features

🎥 Real-time Exercise Tracking

  • 📷 Real-time Pose Estimation using MediaPipe
  • 🎯 18 Built-in Exercises - Full body workout coverage
  • 📊 Form Score System (0-100) with A-F grading
  • 🔄 Automatic Rep Counting with state machine logic
  • 💬 Real-time Form Feedback - Instant correction tips

📹 Video Analysis Mode

  • 🎬 Upload & Analyze Videos - Process pre-recorded workout videos
  • 🦴 Skeleton Overlay - See your pose detection on processed video
  • 📈 Live Statistics Panel - Real-time rep count, form score, and state
  • 🖥️ Processing Terminal - Watch analysis progress with detailed logs
  • 💾 H.264 Video Output - Browser-compatible processed videos with imageio-ffmpeg

👤 User Profile System

  • 📋 Personal Information - Track your fitness journey
  • 🎯 Customizable Goals - Set weekly workout and rep targets
  • 🏅 Achievement System - Unlock badges for milestones
  • 📊 Activity Charts - Visualize your workout history
  • ❤️ Favorite Exercises - Track your most-used exercises
  • ⚙️ Settings - Dark mode, notifications, units preference

📊 Dashboard

  • 📈 Workout Statistics - Total workouts, reps, streaks
  • 📉 Weekly Activity Charts - Visualize your progress
  • 🥧 Exercise Distribution - See which exercises you do most
  • 📋 Recent Workouts - Quick view of latest sessions

⚙️ Extensible Architecture

  • 📝 YAML-based Exercise Definitions - Add new exercises without writing code!
  • 🔀 Three Exercise Types - Standard, Bilateral (left/right), Duration-based
  • 🎨 Customizable Visualization - Colors, highlighted joints per exercise

🏗️ System Architecture

┌─────────────────────────────────────────────────────────────────────┐
│                        FRONTEND (Web UI)                             │
│         index.html ← video stream (MJPEG) ← form feedback            │
└────────────────────────────────┬────────────────────────────────────┘
                                 │ HTTP/WebSocket
┌────────────────────────────────▼────────────────────────────────────┐
│                         BACKEND (Flask)                              │
│                            app.py                                    │
│   • Video capture & streaming                                        │
│   • REST API endpoints                                               │
│   • Session management                                               │
└────────────────────────────────┬────────────────────────────────────┘
                                 │
┌────────────────────────────────▼────────────────────────────────────┐
│                      EXERCISE ENGINE                                 │
│  ┌────────────────┐  ┌────────────────┐  ┌────────────────┐         │
│  │  BaseExercise  │  │    Loader      │  │     Engine     │         │
│  │   (FSM Core)   │  │  (YAML→Obj)    │  │   (Wrapper)    │         │
│  │                │  │                │  │                │         │
│  │ • State Machine│  │ • Parse YAML   │  │ • process_frame│         │
│  │ • Rep Counter  │  │ • Validate     │  │ • draw_overlay │         │
│  │ • Form Score   │  │ • Create Obj   │  │ • draw_score   │         │
│  │ • Feedback     │  │                │  │                │         │
│  └────────────────┘  └────────────────┘  └────────────────┘         │
│         ▲                                                            │
│         │ inheritance                                                │
│  ┌──────┴───────┐  ┌────────────────┐                               │
│  │  Bilateral   │  │   Duration     │                               │
│  │  Exercise    │  │   Exercise     │                               │
│  │ (L/R sides)  │  │ (time-based)   │                               │
│  └──────────────┘  └────────────────┘                               │
└────────────────────────────────┬────────────────────────────────────┘
                                 │ loads
┌────────────────────────────────▼────────────────────────────────────┐
│                    YAML DEFINITIONS (18 exercises)                   │
│  squat.yaml │ push_up.yaml │ plank.yaml │ deadlift.yaml │ ...       │
│                                                                      │
│  Each YAML contains:                                                 │
│  • Angle definitions (landmarks, ranges)                             │
│  • State machine (states, transitions, conditions)                   │
│  • Counter rules (when to count a rep)                               │
│  • Form feedback rules (warnings, priorities)                        │
│  • Tempo guidance (up/down/hold timing)                              │
│  • Visualization config (colors, highlighted joints)                 │
└─────────────────────────────────────────────────────────────────────┘

📊 Form Score System

The app calculates a Form Score (0-100) for each exercise session:

Component Weight Description
Angle Accuracy 40% How close your angles are to ideal
Tempo Compliance 30% Following the recommended speed
Form Feedback 30% Penalty for triggered warnings

Grade Scale

Score Grade Color
90-100 A 🟢 Green
80-89 B 🔵 Blue
70-79 C 🟡 Yellow
60-69 D 🟠 Orange
0-59 F 🔴 Red

📁 Available Exercises (18)

Upper Body

  • 💪 Hammer Curl - Bicep curl with neutral grip
  • 💪 Bicep Curl - Classic bicep exercise (bilateral)
  • 💪 Tricep Dip - Chair/bench dips for triceps
  • 💪 Shoulder Press - Overhead pressing
  • 💪 Lateral Raise - Side delt raises
  • 💪 Push Up - Classic chest exercise

Lower Body

  • 🦵 Squat - Bodyweight squat
  • 🦵 Lunge - Forward lunge (bilateral)
  • 🦵 Side Lunge - Lateral lunge
  • 🦵 Deadlift - Romanian deadlift for hamstrings
  • 🦵 Glute Bridge - Hip bridge for glutes
  • 🦵 Calf Raise - Standing calf raises
  • 🦵 Wall Sit - Isometric hold (duration)

Cardio / Full Body

  • 🔥 Mountain Climber - Dynamic core/cardio
  • 🔥 High Knees - Running in place
  • 🔥 Jumping Jack - Classic cardio move
  • 🔥 Leg Raise - Lying leg raises for abs
  • 🧘 Plank - Core hold (duration)

🚀 Quick Start

1. Clone & Install

git clone https://github.com/yourusername/fitness-trainer-pose-estimation.git
cd fitness-trainer-pose-estimation

# Create virtual environment (optional)
python -m venv venv
source venv/bin/activate  # Linux/Mac
# or: venv\Scripts\activate  # Windows

# Install dependencies
pip install -r requirements.txt

2. Run the Application

python app.py

3. Open in Browser

Navigate to: http://127.0.0.1:5000

➕ Adding New Exercises

No coding required! Just create a YAML file:

Step 1: Create YAML in exercises/definitions/

# exercises/definitions/my_new_exercise.yaml
name: "My New Exercise"
type: "standard"  # or "bilateral" / "duration"
description: "Description of the exercise"

# Define body angles to track
angles:
  primary_angle:
    landmarks: [11, 13, 15]  # MediaPipe landmark IDs
    range: [30, 170]         # Valid angle range

# State machine definition
states:
  - name: "START"
    condition:
      angle: "primary_angle"
      operator: ">"
      value: 150
    next_state: "MIDDLE"
    feedback: "Starting position"
    
  - name: "MIDDLE"
    condition:
      angle: "primary_angle"
      operator: "<"
      value: 60
    next_state: "END"
    feedback: "Good form!"
    
  - name: "END"
    condition:
      angle: "primary_angle"
      operator: ">"
      value: 150
    next_state: "START"
    feedback: "Rep complete!"

# When to count a rep
counter:
  increment_on: "END"

# Form warnings
feedback:
  - name: "form_warning"
    description: "Bad form detected"
    angle: "primary_angle"
    condition:
      operator: "<"
      value: 30
    message: "Don't go too low!"
    priority: 1

# Tempo in seconds
tempo:
  up_seconds: 1.0
  down_seconds: 2.0
  hold_seconds: 0.5

# Display settings
visualization:
  primary_angle: "primary_angle"
  show_angles: ["primary_angle"]
  highlight_landmarks: [11, 13, 15]
  color_scheme: "green"

Step 2: Test

python test_engine.py

Step 3: Use

Restart the app - your exercise is now available!

📍 MediaPipe Landmarks Reference

 0  = NOSE                    
11  = LEFT_SHOULDER       12 = RIGHT_SHOULDER
13  = LEFT_ELBOW          14 = RIGHT_ELBOW
15  = LEFT_WRIST          16 = RIGHT_WRIST
23  = LEFT_HIP            24 = RIGHT_HIP
25  = LEFT_KNEE           26 = RIGHT_KNEE
27  = LEFT_ANKLE          28 = RIGHT_ANKLE
31  = LEFT_FOOT_INDEX     32 = RIGHT_FOOT_INDEX

📂 Project Structure

fitness-trainer-pose-estimation/
├── 📄 app.py                    # Flask application + video streaming
├── 📄 video_processor.py        # Standalone video analysis with skeleton overlay
├── 📄 main.py                   # CLI runner (standalone)
├── 📄 requirements.txt
│
├── 📁 exercises/
│   ├── 📄 base_exercise.py      # FSM engine (BaseExercise, Bilateral, Duration)
│   ├── 📄 loader.py             # YAML loader & validator
│   ├── 📄 engine.py             # High-level API wrapper
│   └── 📁 definitions/          # 🎯 YAML exercise files (18 exercises)
│       ├── squat.yaml
│       ├── push_up.yaml
│       ├── hammer_curl.yaml
│       ├── bicep_curl.yaml
│       ├── tricep_dip.yaml
│       ├── shoulder_press.yaml
│       ├── lateral_raise.yaml
│       ├── lunge.yaml
│       ├── side_lunge.yaml
│       ├── deadlift.yaml
│       ├── glute_bridge.yaml
│       ├── calf_raise.yaml
│       ├── wall_sit.yaml
│       ├── plank.yaml
│       ├── mountain_climber.yaml
│       ├── high_knees.yaml
│       ├── jumping_jack.yaml
│       └── leg_raise.yaml
│
├── 📁 pose_estimation/
│   ├── 📄 estimation.py         # MediaPipe wrapper
│   └── 📄 angle_calculation.py  # Angle math
│
├── 📁 feedback/
│   ├── 📄 indicators.py         # UI components
│   ├── 📄 information.py        # Exercise metadata
│   └── 📄 layout.py
│
├── 📁 db/
│   └── 📄 workout_logger.py     # Workout history logging
│
├── 📁 templates/
│   ├── 📄 index.html            # Home - Real-time exercise tracking
│   ├── 📄 video_analysis.html   # Video upload & analysis page
│   ├── 📄 dashboard.html        # Stats dashboard
│   └── 📄 profile.html          # User profile & settings
│
└── 📁 static/
    ├── 📁 css/
    │   ├── 📄 style.css         # Global styles
    │   ├── 📄 dashboard.css     # Dashboard page styles
    │   ├── 📄 profile.css       # Profile page styles
    │   └── 📄 video_analysis.css # Video analysis styles
    ├── 📁 js/
    │   ├── 📄 script.js         # Main page JavaScript
    │   ├── 📄 dashboard.js      # Dashboard functionality
    │   ├── 📄 profile.js        # Profile page functionality
    │   └── 📄 video_analysis.js # Video analysis functionality
    └── 📁 images/

🔌 API Endpoints

Endpoint Method Description
/ GET Home page - Real-time exercise tracking
/video_analysis GET Video upload & analysis page
/dashboard GET Workout statistics dashboard
/profile GET User profile & settings
/video_feed GET MJPEG video stream
/start_exercise POST Start tracking an exercise
/stop_exercise POST Stop current exercise
/get_status GET Get current rep count & form score
/exercises GET List all available exercises
/api/video/upload POST Upload video for analysis
/api/video/status/<id> GET Get video analysis status
/api/video/processed/<id> GET Download processed video
/api/profile/update POST Update user profile

🖼️ Screenshots

Home - Real-time Tracking

Real-time pose estimation with skeleton overlay and form feedback.

Video Analysis

Upload videos, analyze with skeleton overlay, and download processed results.

Profile

Personal stats, achievements, goals tracking, and customizable settings.

🛠️ Technologies

  • Flask - Web framework
  • OpenCV - Computer vision & video processing
  • MediaPipe - Google's pose estimation model
  • imageio-ffmpeg - H.264 video encoding for browser compatibility
  • PyYAML - Exercise definition parsing
  • Chart.js - Interactive charts for dashboard
  • HTML/CSS/JS - Modern responsive frontend

🤝 Contributing

  1. Fork the repository
  2. Create a feature branch (git checkout -b feature/new-exercise)
  3. Add your YAML exercise definition
  4. Run tests (python test_engine.py)
  5. Commit changes (git commit -am 'Add new exercise')
  6. Push to branch (git push origin feature/new-exercise)
  7. Create Pull Request

📜 License

MIT License - see LICENSE file for details.

🙏 Acknowledgments

Made with ❤️ for fitness enthusiasts

About

An AI-powered fitness tracker that uses real-time pose estimation to count reps, monitor form, and provide instant feedback for exercises like squats, push-ups, and bicep curls. Designed for accuracy, motivation, and adaptability

Topics

opencv computer-vision image-processing artificial-intelligence fitness-app pose-estimation fitness-tracker computer-vision-tools computer-vision-opencv pose-detection fitness-trainer

Resources

Readme

License

MIT license
Activity

Stars

65 stars

Watchers

1 watching

Forks

25 forks
Report repository

Releases

No releases published

Packages

 
 
 

Contributors

Languages