Validation of Microsoft Kinect for Use in Detecting Balance Impairment in ACL Repaired Patients

Project Enquiry:

Fields with * are mandatory


The anterior cruciate ligament (ACL) is a critical part of normal knee function. With as many as 250,000 injuries every year, most of which occur through little or no contact, ACL reconstruction and rehabilitation are extremely important and relevant topics. One way to test a patient’s rehabilitation from ACL reconstruction is through the Star Excursion Balance Test (SEBT).

A subject’s reach distance in the SEBT can be determined using motion capture systems.  However, state of the art motion capture systems present some difficulties, such as cost and complexity of use. The Microsoft Kinect©, an inexpensive,  markerless motion capture system, was predicted to solve the issues facing modern motion  capture systems, while portraying acceptable levels of accuracy.

It was also predicted that  subjects who had undergone ACL reconstruction within the previous three months would not  perform as well as controls in the SEBT.It was concluded that in most directions of the  SEBT, the Kinect© performed within an acceptable range for use by clinicians in determining a  patient’s advancement through rehabilitation. It was also concluded that the pathological subjects did not perform as well as controls. Future directions for this study include testing more pathological subjects and controls, as well as determining ways to capture more difficult directions of the SEBT.


Anatomy and Physiology:

The anterior cruciate ligament (ACL) is named so because it crosses and attaches to the tibia anterior to the posterior cruciate ligament, specifically anterior and lateral to the medial intercondylar tubercle. Climbing to the femur, it attaches at the posteromedial surface of the lateral femoral condyle. As it does this, the ACL twists on itself approximately 90°, creating a slight lateral spiral facilitated by its bone attachments (Arnoczky, 1983; Goldblatt & Richmond, 2003).

Anatomy of the knee, A) Anterior viewin full flexion B) Posterior viewin extension.(Drake et al., 2014).

Anatomy of the knee, A) Anterior viewin full flexion B) Posterior viewin extension.(Drake et al., 2014).


Motion Capture Technology:

The analysis of biomechanical movements is often used in rehabilitation and treatment as it provides an opportunity to obtain information that is difficult or impossible to acquire otherwise (Pfister, West, Bronner, & Noah, 2014). While kinematics can provide information on dynamic joint motion, kinetics is of primary concern when dealing with motor impairments and their responses (Asfour & Eltoukhy, 2011). To this end, motion capture systems such as Vicon Nexus Motion Capture System (Oxford Metrics, UK), a commercially available and validated gait analysis package, provide invaluable kinetic data to assist researchers in studying gait, balance and other biomechanical tasks (Pfister et al., 2014).

Marker Placement for Full Body Motion Capture(Marker Placement Protocols).

Marker Placement for Full Body Motion Capture(Marker Placement Protocols).


Stated Goals:

With the prevalence of ACL injuries and rehabilitation, it is important to be able to measure the progress of a patient through the rehabilitation process to confirm that the assigned protocol is achieving the desired goals. Using the SEBT recorded by a motion capture system would allow clinicians the ability to confirm that a patient’s rehabilitation protocol is working as expected.

As traveling to a lab can present inconveniences to the patient, a portable motion capture system that can be used in a clinician’s office would allow the patient to perform the SEBT at the time of their meeting with a clinician, and would give the clinician the information to determine if the patient is responding well to the rehabilitation protocols.

Acquiring Kinect© Data:

To be able to use the Kinect© and obtain data, the Software Development Kit 2.0 from Microsoft was used in conjunction with MATLAB R2016a (The Mathworks, Inc., Natick, MA). A variety of programs were written in MATLAB to be able to perform the tasks needed to acquire data from the Kinect© and compare its output to the Vicon data.


Validation of Kinect© versus Vicon:

For the validation, the Star Excursion Balance Test data capture by the Kinect© and Vicon was compared at 10-90% of the total movement, in increments of 10%. Afterwards, the error and standard deviation of all the points were taken. This would allow for an accurate comparison of the ability of the Kinect©system to track the Star Excursion Balance Test against a modern marker-based system such as Vicon.


Validation of Kinect© versus Vicon:

Ten control subjects were recorded performing the Star Excursion Balance Test by both the Kinect© and Vicon. The average of all data in each direction is shown below.

ACL Patients Versus Controls:

Three pathological subjects were included in the study. The first subject was recorded three weeks post surgery, which included ACL and meniscal repair. They did not feel comfortable performing the Star Excursion Balance Test on their injured leg, but was able to perform the test on their uninjured leg.


Future directions of this study include recruiting more control subjects to better and more accurately validate the Kinect© with the Vicon system. More data would allow for more fine-tuning of the Kinect©, as well as allow for the formulation of a regression equation. More trials may also be beneficial to potentially capture some directions that have proven difficult to capture, such as the Medial direction.

In regards to comparing ACL injured subjects to controls, more pathological subjects should be included in the study to be able to expand upon the current data so that more relevant statistics can be run.

Also, just as the Y test has been developed for the study of ankle instability, a similar test could be formulated that reduces the number of directions needed to determine the progress of rehabilitation of the subject (Gribble et al., 2012). If it could be determined that performance in a smaller number of directions correlates with a subject’s performance for all direction in the SEBT, it is possible that some directions that are difficult to capture may be eliminated.

Source: University of Miami
Author: Andre Alvarez

Download Project

>> Best Matlab Project Ideas for Engineering Students with Full Project Materials

Project Enquiry:

Fields with * are mandatory

Leave a Comment

Your email address will not be published. Required fields are marked *