3D Motion Capture with a monocular lens system technique that uses artificial intelligence to analyze human movements from a single camera.
The system consists of multiple Neural Networks that are trained using data from marker and markerless cameras, force plates and sensors.
The Neural Networks are optimized for CPU and GPU on IOS and Android devices, so they can run efficiently on mobile platforms. The system captures body points using a single camera and uses the AI to predict body points from occluded camera views, such as when the limbs are hidden behind the torso.
The system then calculates the musculoskeletal movement of joints in the frontal, sagittal and transverse planes with six degrees of freedom, which means it can measure the rotation and translation of each joint in three dimensions.
This way, the system can provide accurate and detailed information about human motion from a single camera perspective.