Understanding how your video becomes a detailed movement report can help you better interpret your results. In this article, we’ll explain how the 3MotionAI engine transforms a simple video into science-level biomechanical insights.
It All Starts With a Video
All you need is a video captured using:
Our 3MotionAI Pro app, or
Any smartphone or video capture device
Once your video is uploaded, our processing engine begins analyzing the movement automatically.
The 3-Layer Processing Engine
At the core of 3MotionAI is our 3DNeuroNET Engine — the system that converts standard video into measurable, real-world movement data. Here’s how it works:
Layer 1: Computer Vision Tracking
The first step is computer vision. Our system:
Detects and tracks 19 major body joints in 2D
Analyzes movement frame by frame
You can think of this as a moving digital stick figure that follows the person’s motion over time. This layer creates the foundation for everything that follows.
Layer 2: 3D Reconstruction
Next, our proprietary AI model transforms that 2D information into a detailed 3D skeletal model. What this means:
2D pixel data becomes real-world measurements (e.g., metres, degrees)
Depth (in/out of frame) is accurately reconstructed
The expands to 53 anatomical landmarks (in 3D)
Unlike basic vision tools that struggle with depth estimates, our engine is designed to provide highly accurate 3D tracking compared to traditional motion capture methods.
Layer 3: Task-Specific Metrics & Insights
Once the 3D model is built, the system calculates meaningful metrics based on the assessment type you selected. Depending on the task, this may include:
Joint angles
Reach distances
Postural patterns
Risk indicators
Other task-specific movement metrics
These calculations form the foundation of the results and visualizations you see in your final report.
Why This Matters for Your Reports
Understanding this process helps you:
Interpret joint angle measurements more confidently
Recognize how depth and 3D motion are captured
Trust the consistency of your analysis
Better compare results across assessments
All of the metrics in your report are built on this processing pipeline. See the video below from our Video Education Series on our 3DNeuroNET engine.