Latest News

Digital Camera and Lenses

Theia’s Markerless Motion Capture: First Ever Validated as Accurate as Marker-based Systems!

Posted 2021-08-04

Yesterday, a third scientific validation study of Theia Markerless’s AI-driven motion capture technology was published in the Journal of Biomechanics. Titled “Concurrent Assessment of Gait Kinematics Using Marker-based and Markerless Motion Capture” this study is part of a series that has validated the accuracy of Theia’s markerless motion capture system against the traditional, industry standard of marker-based, infrared motion capture. This is the first, ever, third party verification that a markerless motion capture technology is as accurate, repeatable, and reliable as marker-based motion capture with regards to measuring and quantifying human movement by breaking it down to its component forces and angles.

“We are very excited to see this study published,” said Scott Selbie, CEO and co-founder of Theia Markerless, “It confirms our belief that markerless technology is the future of motion capture. Our data-driven approach will only grow in utility and versatility in the coming months and years, and we will continue to shift the needle towards accessible and accurate data on human movement.” 

The previous studies in this series focused on the inter-session repeatability of markerless motion capture and whether it can be used to assess gait (how people walk) as well as marker-based systems. Combined, these studies show that Theia’s motion capture is a viable alternative to marker-based systems, even when measured using the most rigorous scientific standards of biomotion.

The previously published studies in this series are:

Adopting Markerless

Posted 2021-07-08

Dr. Kevin Deluzio and Dr. Elise Laende discuss their use of Theia3D in the lab.

Check out the latest Theia3D validation studies!

Updated 2021-08-23

Concurrent Assessment of Gait Kinematics Using Marker-based and Markerless Motion Capture

Kanko, R., Laende, E., Davis, E., Selbie, S., & Deluzio, K. (2021, Oct 11). https://doi.org/10.1016/j.jbiomech.2021.110665
 
 

Assessment of spatiotemporal gait parameters using a deep learning algorithm-based markerless motion capture system 

Kanko, R., Laende, E., Strutzenberger, G., Brown, M., Selbie, S., DePaul, V., Scott, S., & Deluzio, K. (2021, June 9). https://doi.org/10.1016/j.jbiomech.2021.110414

(Archived preprint https://engrxiv.org/j4rbg/)

Inter-session repeatability of Theia3D markerless motion capture gait kinematics

Kanko, R., Laende, E., Selbie, S., & Deluzio, K. (2020, June 23). https://doi.org/10.1016/j.jbiomech.2021.110422

(Archived preprint https://www.biorxiv.org/content/10.1101/2020.06.23.155358v1)

Latest Research on Theia3D

The basketball equipment

Every year new research articles are published using Theia3D. In order to synthesize these publications we have provided links to the original articles below. If you would like a full description of the abstracts please email us directly and our team can provide a copy of the latest Theia3D Research Compendium. 

 

To learn more about Theia3D, click here to book a demo. 

Why Markerless?

Simplify Setup:

Traditional motion capture requires subjects to wear tight-fitting clothes and dozens of reflective markers that must be applied by a trained operator. This setup process is time-consuming, inconvenient, and often takes longer than the collection itself. With markerless tracking there are no markers, no special clothes, and almost no setup required.

Collect Anywhere:

Collect athletes on the court or on the field, assess patients in the clinic, and measure movement in the real world. Unlike traditional motion capture systems, markerless tracking is not limited to a specialized laboratory or studio space so you can study movement where it happens. 

Collect Real Movements:

Athletes perform best in-game and people behave naturally in familiar environments. Using markerless motion capture, you can record real movements in the the real world without the artifact introduced by the laboratory environment, clothing, and markers required by traditional motion capture systems.

Collect More Data:

Little to no setup means you can collect more subjects in a day. Consistent and repeatable results allow you to perform multi-center studies. No markers and no lab so you can collect anyone, anywhere. With markerless motion capture you can do more with less.

Why Theia3D?

Biomechanics Focus:

Theia3D isn’t just another markerless tracking solution – it’s markerless tracking for biomechanics. Every aspect of Theia3D, from the 2D features tracked, to the segment definitions and joint constraints, has been selected to ensure biomechanical accuracy and relevance of the results and to follow standard practice and conventions in the field.

Accurate and General Solution:

Theia3D distinguishes itself from the other markerless tracking solutions through its ability to provide a highly accurate and generalized solution to markerless motion capture.

Theia3D uses rigid-body tracking and inverse kinematics to fully solve the 3D pose of skeletal segments. Compared to edge-detection and hull-carving techniques, this approach is better for resolving segment rotation to provide an accurate and complete pose estimate.

Theia3D is a general solution – it can track a wide range of movements in almost any environment. This is a critical distinction from other specialized markerless tracking solutions that are limited to tracking a specific activity (such as pitching or batting) and cannot be used as a general solution.

Multi-Person Tracking and Automatic Model Creation:

When multiple people are visible in the camera views, they are automatically identified and tracked in each view and in 3D. Models are automatically created and scaled for each identified person using parameters inferred from the videos.

Automatic and Batch Processing:

The entire analysis pipeline can be executed automatically with no input from the user. This enables efficient batch processing of multiple files. A data maintenance tool is include to automatically organize data for batch analysis.

View and Export Results:

Tracked people are rendered in an interactive 3D scene and in 2D video overlays for each of the cameras that are easily exported as video files to visualize results and create presentation materials. Inverse kinematics results can be displayed in the built-in joint and segment angle viewer or exported to a C3D file for further analysis. Export to common animation formats is coming soon.

Integration with Visual3D:

Exported C3D files can be opened in Visual3D with automatic model creation for enhanced analysis and reporting. Automatic reports are also available. Visual3D is the premier biomechanics software for measuring movement and offers an unparalleled suite of analysis and reporting tools.

How Does it Work?

There are three steps to markerless motion capture with Theia3D:

1. Collect the data.

Theia3D requires synchronous, calibrated, video data. There are a number of camera systems capable of collecting this data including the Sony RX0 II camera bundle that can be purchased with the software. Other capable camera systems are outlined in the Camera System Options article.

2. Process the data.

Theia3D identifies and tracks the people in the video views and then computes the 3D position and orientation of each segment of each tracked person.

3. Use the data.

Theia3D exports a C3D file that can be opened with Visual3D for further analysis and reporting. Export to common animation formats is coming soon.

FAQs

Q: Can I use my own cameras?

A: Yes you can, however the cameras must collect fully synchronized, high quality video data. We only support data collected using specific camera systems that we have verified the quality and synchronization of the captured videos. So, while you can use your own camera system, if it is not on our list of verified equipment it is your responsibility to ensure the system meets the requirements of Theia3D, and we will not be able to provide support to any issues related to data collection or quality.

Q: What is the maximum sampling rate? How about data capture length?

A: The maximum sampling rate and capture length are limitations of the cameras used and thus are different for every camera system. Longer captures will take longer to analyze in Theia3D. 

Q: Can the system be used outdoors?  (How about low-light situations?)

A: Yes. This system can be used outdoors and will perform well in any environment as long as the lighting is such that the person is clearly visible, and key features, such as the hands and feet, are easily distinguishable in the video. 

Q: What are the limitations on the capture volume?

A: The capture volume size will be limited by the number of cameras. Larger volumes and volumes of more complex shape require more cameras. 

Q: How long does the analysis take?

A: It depends on the computer, the number of GPUs available, the number of cameras, and the capture frame rate. A typical collection of 10 seconds, collected at 40 frames per second, for 8 cameras will take around 1 minute to process on a single high-end graphics card. An Nvidia RTX 2080Ti or better is the recommended graphics card.

Q: Can I include force data for kinetics?

A: Only if your camera system supports the synchronized collection of external devices. The Qualisys Miqus camera system, that has been verified to work with Theia3D, has hardware options for collecting synchronized data from force plates, instrumented treadmills, EMG, and marker-based cameras.

Q: How fixed are the models? Does it support non-standard subjects like prosthetics, animals, or a multi-segmented foot? 

A: Currently, Theia3D supports a single model. That being said, abnormal movements and pathological gait can be tracked using this software. 

Q: Can the processing be made automatic?

A: Yes, after the system is setup, the 3D analysis process can be fully automated for single trial or batch processing. User input is only required to start the process.

Q: Can I export data for reporting? Is there built in reporting?

A: Report generation and biomechanical analysis functionality is not built into Theia3D. To generate reports, we recommend using Visual3D (which is included in our packaged offerings) as it can automatically detect the data exported by Theia3D and build the required models. However, we also export C3D files that can be used in other software.

Q: Does Theia3D support animation or gaming development?

A: At this time Theia3D only exports C3D formatted files and AVI video files with segment/skeletal overlays applied. Future versions may export FBX formatted files.

Q: Does this work on multiple people or just one?

A: Yes. Theia3D identifies, tracks, and analyzes every person that satisfies the visibility requirements. 

Q: Do I need any other software other than Theia3D?

A: Yes. You need software to run the cameras and collect the video data. Theia3D processes this video data to produce 3D kinematic data (segments and rotation matrices) that is ready for analysis. Additional analysis or 3D processing software (such as Visual3D) can be used to interpret and report the 3D kinematic data.

Q: Has Theia3D been validated?

A: Recent studies have demonstrated the concurrent validity of Theia3D with existing marker based motion capture systems when measuring spatiotemporal gait parameters (https://engrxiv.org/j4rbg/), shown that the gait kinematics measured with Theia3D are more reliable than those measured with existing marker based systems (https://biorxiv.org/cgi/content/short/2020.06.23.155358v1), and shown that joint angles for the lower extremity measured with Theia3D are within the error range of uncertainty for marker based motion capture (https://biorxiv.org/content/10.1101/2020.12.10.420075v1).

Q: How accurate are the results compared to other markerless systems?

A: Because of our dedication to data quality for biomechanics, Theia3D is state of the art in terms of accuracy. To our knowledge, no markerless system has been validated to the degree of Theia3D.

Markerless Hardware Options

Video Equipment in Action

Camera System

Theia3D requires a minimum of 6 synchronized video cameras to track movement. However, for accurate tracking in a modest volume a minimum of 8 cameras is recommended. Larger or more complex volumes, more complex movements, or additional people in the volume may require additional cameras. Please see the Camera System Requirements wiki page for more information.

Theia3D is not limited to a specific camera system and in general will work with any system that provides synchronized video files and meets the other Camera System Requirements. There are several camera systems available that can meet these requirements; however, it is important to verify the accuracy of their synchronization. To date, the systems that we have tested and verified to have good performance with Theia3D are:

  • Sony RX0 II (www.sony.com) – These are the cameras included in the complete markerless motion capture package that can be purchased directly from Theia. Note that synchronous capture with these cameras requires additional hardware such as control boxes and a network switch (all required hardware is included in the markerless motion capture package).

In addition to the independent camera systems we have tested above, our solution providers offer Theia3D packages with different camera options and the ability to synchronize other hardware such as force plates, an instrumented treadmill, EMG sensors, or traditional marker based cameras with the markerless system. These solution providers include:

Computer

Processing data with Theia3D requires a desktop computer with a CUDA capable graphics card, multi-core processor, and reasonable amount of RAM. The minimum and recommended system specifications are outlined below. Note that any use of Theia3D on a machine that does not meet these minimum requirements, or on a laptop is not supported.

  • GPU: One or more CUDA capable graphics cards. NVIDIA RTX 2080,  NVIDIA RTX 2080 Super, NVIDIA RTX 2080Ti, NVIDIA RTX 3080, and NVIDIA RTX 3090 are supported.
  • RAM: Minimum 24 GB. Recommended 32 GB.
  • CPU: Minimum quad-core i7 processor or equivalent. Recommended octa-core i9 processor.
  • Updated graphics card driver. The program will not run if the graphics driver is not up to date.