The ultimate computer vision frontend for biomechanical analysis. Track path, velocity, and alignment in real-time.
Technical Architecture
CORE_SYSTEMS_OVERVIEW
Path Tracking
Tracking the verticality of the load (Bar Path) to minimize moment arms and sheer force.
Angular Velocity
Detecting "sticking points" via joint speed analysis to identify compensation patterns.
Segment Alignment
Ensuring joint centration to prevent common energy leaks like knee valgus.
Development Roadmap
SYSTEM_EVOLUTION_PATH
Computer Vision Core
Implement MediaPipe Pose Landmarker for 30fps real-time skeletal tracking. Establish coordinate normalization to handle different user heights and camera distances.
Computer Vision Core
Implement MediaPipe Pose Landmarker for 30fps real-time skeletal tracking. Establish coordinate normalization to handle different user heights and camera distances.
The Biomechanical Engine
Develop the state machine for Temporal Analysis (Eccentric/Concentric detection). Implement joint angle calculation (e.g., Hip Flexion = vector(Hip, Knee) vs vector(Hip, Shoulder)).
The Biomechanical Engine
Develop the state machine for Temporal Analysis (Eccentric/Concentric detection). Implement joint angle calculation (e.g., Hip Flexion = vector(Hip, Knee) vs vector(Hip, Shoulder)).
Real-Time Feedback Loop
Integrate Text-to-Speech API for sub-200ms audio corrections. Build the 'Quality Filter' to discard reps that don't meet IPF depth standards.
Real-Time Feedback Loop
Integrate Text-to-Speech API for sub-200ms audio corrections. Build the 'Quality Filter' to discard reps that don't meet IPF depth standards.
Gemini Context Integration
Connect the 'Gold Standard' feature. Upload set kinematics to Gemini-1.5-Pro to compare against elite athlete templates and generate the 'PhD Audit'.
Gemini Context Integration
Connect the 'Gold Standard' feature. Upload set kinematics to Gemini-1.5-Pro to compare against elite athlete templates and generate the 'PhD Audit'.
Predictive Analytics
Use historical velocity loss data to predict 1RM and suggest autoregulated load adjustments for the next session.
Predictive Analytics
Use historical velocity loss data to predict 1RM and suggest autoregulated load adjustments for the next session.
Kinematic Landmarks
MEDIAPIPE_POSE_MAPPING
Squat
Hip
To track depth relative to knees.
Knee
To detect valgus/varus collapse.
Ankle
To measure shin angle and mobility.
Shoulder
To monitor torso inclination and bar path.
Bench Press
Wrist
To ensure stacked joints under the bar.
Elbow
To track flare angle relative to torso.
Shoulder
To detect scapular retraction loss.