opencap-org/opencap-core

What it solves

OpenCap Core provides a way to estimate 3D marker positions and human movement kinematics (such as joint angles) using only two or more smartphone videos, removing the need for expensive, marker-based motion capture systems.

How it works

The system processes multiple video streams to estimate 3D human movement. It integrates with OpenPose for pose estimation and OpenSim for kinematic analysis. Users can either use a cloud-based web application for automatic processing or run the pipeline locally on their own hardware to customize settings, reprocess data with higher accuracy, or use videos from other synchronous sources.

Who it’s for

It is primarily designed for academic researchers and clinicians who need to analyze human movement dynamics and kinematics without specialized laboratory equipment.

Highlights

• Multi-video input: Estimates 3D positions from two or more video sources.
• OpenSim Integration: Outputs data in OpenSim format, allowing for further kinetic (force) calculations.
• Flexible Deployment: Available as a cloud service for academic research or as a local installation for advanced users.
• High-Accuracy Options: Supports various pose estimation settings depending on available GPU memory.