brondijk.xyz

🏋️ Overview

Squat Coach is a computer vision application that helps users perfect their squat form. Using a webcam, it tracks body movements in real-time, provides feedback on form, and counts reps with proper technique.

✨ Features

• Real-time body pose tracking using webcam
• Automatic squat rep counting
• Form analysis with feedback (depth, knee alignment, back angle)
• Progress tracking over time
• Customizable workout plans
• Audio cues for form correction

🛠️ Implementation

Squat Coach uses MediaPipe's pose estimation model to identify 33 key body landmarks in real-time. Custom algorithms analyze the relationships between these points to determine:

• Hip angle (for squat depth)
• Knee alignment (to prevent valgus)
• Back angle (to maintain proper posture)
• Rep detection (using state machine logic)

The application provides visual feedback by overlaying guidelines and indicators on the video feed, along with audio cues for form corrections.

def analyze_squat_form(landmarks):
    # Calculate hip angle
    hip_angle = calculate_angle(
        landmarks[LANDMARK_MAP['left_shoulder']],
        landmarks[LANDMARK_MAP['left_hip']],
        landmarks[LANDMARK_MAP['left_knee']]
    )
    
    # Calculate knee alignment
    knee_alignment = check_knee_alignment(
        landmarks[LANDMARK_MAP['left_hip']],
        landmarks[LANDMARK_MAP['left_knee']],
        landmarks[LANDMARK_MAP['left_ankle']]
    )
    
    # Determine squat depth
    if hip_angle < 90:
        depth = "DEEP"
    elif hip_angle < 110:
        depth = "PARALLEL"
    else:
        depth = "SHALLOW"
        
    return {
        "depth": depth,
        "knee_alignment": knee_alignment,
        "hip_angle": hip_angle
    }

🔮 Future Improvements

• Support for additional exercise types (deadlifts, lunges, etc.)
• Mobile app version for iOS and Android
• Cloud-based progress tracking and analytics
• Integration with fitness wearables for heart rate monitoring
• Personalized coaching based on user progress