LOOOK: Live MOOC Orchestration using Kinect

Encadrants : 

Occurrences : 


Nombre d'étudiants minimum: 


Nombre d'étudiants maximum: 


Nombre d'instances : 


Massive Open Online Courses (MOOCs) are online courses accessible via a web interface. Traditionally they consist of digital material collections on a topic, such as slides, notes etc. and pre-recorded video lectures. In order to reduce the time required for a course to be accessible, we propose the live streaming of the lectures. This approach has many challenges – like the timing of the slides according to the lecture progress, and other functions that are available by post-production techniques on the recorded material.

Kinect is a multimodal input device, originally created as a peripheral of Microsoft’s Xbox. Due to the fact that it can provide Video, Audio and Depth information of the scene, its use has exceeded the limits of gaming and finds applications in a plethora of topics. We propose using the capabilities of Kinect, to overcome the limitations of a live classroom, and instead offer an even more immerse and extensible classroom environment.

In the typical scenario we target, a teacher is giving a lecture in front of a Kinect camera that records Audio, Video and Depth, a server is encoding the captured data, and then transmits it alongside other information (slides, captions etc), to a remote browser-based client. Then, the web application is responsible for extracting the data, synchronize, and use them in order to produce the final output. Both the server and the client applications will be designed and implemented by the PAF group, which is also expected to consider MOOC features (live slides, video chapters, captions etc).

For the server side, the C/C++ based application will parse the Kinect input data using Kinect SDK (Windows), encode them via FFMpeg and package them with GPAC. Then, the data will be transmitted to the web-based client, over HTTP Streaming. The browser application will be in Javascript and the final output will be rendered in the browser with WebGL. Other tools of trade that might be used are JavaScript 3D and Game Engines (e.g. three.js, Phaser) and/or Artificial Vision (e.g. OpenCV) libraries.


The working languages for this project are French and English [the poster can be either in English or French]

Keywords: Kinect, Distributed Systems, MOOC, Human-Computer Interaction

Covered Topics:

  • Multimedia
  • Programming
  • Computer Networks
  • Web Technologies