Extraction of spatial information for low-bandwidth telerehabilitation applications

1
Extraction of spatial information for
low-bandwidth telerehabilitationapplications

• Kok Kiong Tan, PhD Arun Shankar Narayanan
  Choon Huat Koh, PhD Kevin Caves Helen Hoenig,
  MD, PhD

2

• Aim
• Introduce innovative ways to use information
  about the depth of selected objects with
  2-dimensional (2D) video for telerehabilitation
  applications.
• Relevance
• Rehabilitation assessments may require spatial
  information to accurately assess patients.
• Sending 3-dimensional (3D) video data over
  low-bandwidth networks is extremely challenging.

3
Method

• Used Microsoft Kinect sensor together with
  opensource color detection and tracking library
  package to extract desired depth data from image.
  
• Plotted extracted depth data on graph that was
  then displayed alongside images to allow sensible
  conclusions about patients condition.
• Presented examples of rehabilitation assessments
  for which spatial information is required as well
  as solutions to leverage the proposed approach.

4
Results
Example Tracking a single object.
Example Tracking body joints in cane height
assessment.
5
Conclusion

• Effectiveness was shown for a framework to
  extract information about depth and transmit it
  with 2D images for use in rehabilitation.
• This framework should prove very useful for
  providing information about spatial resolution
  when limited network bandwidth hinders
  conventional 3D video transmission.