With this website we hope to present our ideas, thoughts and findings on the use of depth cameras (such as the Microsoft Kinect) in biomechanics. As well as being a home for our own case studies, we will provide free access to our data and useful programs we have written.
Most recent changes:
20/10/2015 Matlab viewer version 3.1: Minor bug fixes
24/1/2014 New post: Details of our depth distortion correction technique and associated paper
30/12/2013 Matlab viewer version 3: Minor bug fixes
29/12/2013 New menus added to website (content will be added soon)
23/12/2013 Badminton tracking report added to website: Badminton Tracking report
15/4/2013 Matlab viewer version 2: New features, the software now accepts IPIsoft version 2 files
17/1/2013 New post: How sensitive is the Kinect’s sensor to time/temperature
We have identified three areas of use:
- Segment tracking:
Depth cameras such as the Microsoft Kinect have automatic segment tracking capabilities. While the original intention of such functionality was gesture-based computer control, there are obvious applications in a biomechanics context.
- Segment/Body scanning:
A depth camera usually returns a dense cloud of 3D points. A point cloud can be used to form models and volumes of real objects. There is potential to use this for scanning the body to obtain estimations of volume/shape etc.
- Blob tracking:
The 3D information given by a depth camera can be used on a larger scale for tracking moving objects (such as a badminton player on court). This could give information such as position, displacement, velocity etc. automatically and instantly. While different types of depth cameras have different ranges, this application can be used at a variety of scales.
What is a depth camera?
Quite simply, it is a camera which can ‘see’ (and measure) in three dimensions. We’ve written a short page describing the operation of the Microsoft Kinect depth camera here.
The Microsoft Kinect
Arguably the Microsoft Kinect is the consumer device which really kick-started widespread use of depth cameras in academic research. It’s low price and versatility made it the de facto choice for many exploring the area. Needless to say all of our current investigations into depth cameras and biomechanics have been performed using the Microsoft Kinect.
The feasibility of the Microsoft Kinect for use in Biomechanics
We were awarded funding through the EPSRC and Sheffield Hallam University’s Engineering for life project. Much of the work on this website stems from these initial investigations. Our studies focused on the accuracy of Kinect in three distinct areas: segment scanning, person tracking and body segment tracking.
A full report on our findings can be downloaded in pdf format Kinect Project report.
If you’d like access to the data we’ve captured in our investigations into depthbiomechanics, visit this page. We’ve tried to group the data according to the three distinct section discussed above
During the course of our investigations into the use of depth cameras in biomechanics we’ve written software to assist us. Sometimes we’ve thought that a particular program would be of use to the biomechanics community. We make these programs freely available and we present them on this software page.