Week 11

Kinect: Motion Tracking + OpenCV

Class Overview

Lecture 1

  • Assignment review
  • Skeleton Tracking from Kinect SDK (Windows)
  • Methods of getting the coordinates of joints with Skeleton Tracking
  • Motion tracking with point cloud
  • Methods of getting the center coordinates of an area detected in depth range
  • Methods of getting the closest coordinates of an area detected in depth range
  • Compare them to other tracking methods

Lecture 2

  • Discussion on how to improve the sensing quality with Kinect
  • Introduction to OpenCV library
  • Review useful examples from OpenCV library
  • Discussion on how to remove noise by using dilate() & erode() functions from OpenCV
  • Discussion on how to use Contour Tracking and why it is useful


  • Prepare Final Project Concept Presentation.
  • Create a simple interaction with Kinect
    • Please feel free to reach out to the instructor if you struggle with coming up with ideas.
    • The exercise should be uploaded to Kinetic Interfaces google drive folder by midnight every Saturday so that the instructor has enough time to review your work. You should upload not only your source code but also all of your media assets, such as image, sound and video so that the instructor is able to execute the program.
    • Additionally, your work process should be documented on the documentation blog. Your documentation needs to include sufficient amount of screenshots and video captures along with descriptions.