An innovative iOS application that provides real-time exercise form analysis using advanced computer vision and machine learning technology. Formify makes professional-grade fitness guidance accessible to everyone through their mobile device.

Formify leverages MoveNet's pose estimation capabilities combined with Apple's action classification framework to deliver immediate, personalized feedback on exercise technique. The system analyzes your form in real-time, helping you exercise more safely and effectively without needing a personal trainer.

Key Statistics:

• 94% pose detection accuracy across diverse body types
• 95% exercise classification accuracy for 12 common exercises
• 87% agreement with professional trainer assessments
• Sub-200ms latency for real-time feedback

• Accurate Pose Detection: Identifies 17 key body points at 20+ frames per second
• Exercise Recognition: Automatically detects 12 common strength training exercises
• Form Evaluation: Analyzes joint angles, body alignment, and movement patterns
• Instant Feedback: Color-coded visual indicators and directional guidance

• 100% On-Device Processing: All analysis happens locally on your iPhone
• No Data Transmission: Your workout data never leaves your device
• No Internet Required: Works perfectly offline

• Intuitive Setup: Get started in under 5 minutes
• Visual Guidance: Clear, actionable feedback during workouts
• Progress Tracking: Monitor improvements over time
• Post-Workout Summaries: Detailed analysis and recommendations

• Squats
• Deadlifts
• Lunges
• Pushups
• Planks
• Shoulder Press
• Bicep Curls
• Bench Press
• And 4 more...

• MoveNet Architecture: Optimized for mobile deployment with 65% size reduction
• VTON HD Dataset: Trained on high-quality pose data for exercise-specific accuracy
• Core ML Integration: Efficient on-device machine learning
• SwiftUI Interface: Modern, responsive user experience
• Battery Efficient: Approximately 12% drain per hour of use

This project was developed as a capstone project at VIT (Vellore Institute of Technology) by:

• Saket Balabhadruni (21BKT0001)
• Shamya Haria (21BKT0081)
• Aryaman Bhatnagar (21BKT0185)

Under the supervision of Prof. Poornima N
School of Computer Science and Engineering (SCOPE)

Patent Filed - VIT IPR&TT Cell has filed a patent application for the innovative technical approaches developed in this project.

• Home Fitness: Professional guidance for home workouts
• Injury Prevention: Detect risky form errors before they cause injury
• Skill Development: Progressive feedback for mastering proper technique
• Physical Therapy: Monitor rehabilitation exercise compliance
• Athletic Training: Optimize performance through form refinement

• Machine Learning: TensorFlow, Core ML, MoveNet
• iOS Development: Swift, SwiftUI, Vision Framework
• Computer Vision: Pose Estimation, Temporal Pattern Recognition
• Deployment: iOS 14.0+, iPhone XS or newer

• Multi-person tracking for group workouts
• Expanded exercise library (30+ exercises)
• Integration with Apple Health and wearable devices
• 3D pose estimation using depth sensors
• Personalized training programs
• Social features and community challenges

Formify democratizes access to professional-grade exercise coaching, addressing the critical need for proper form guidance that affects millions of fitness enthusiasts worldwide. By preventing injuries and improving workout effectiveness, it contributes to better health outcomes and fitness adherence.

For inquiries about this project, please contact through VIT's official channels.

Institution: Vellore Institute of Technology (VIT)
School: Computer Science and Engineering (SCOPE)
Year: 2025

This is a private academic project. All repositories within this organization are private and not available for public access or contribution.