AI-Based Media Player Development

Dash.js is a production quality multimedia player, developed as an initiative of the DASH Industry Forum. It can play back MPEG-DASH content using client-side JavaScript libraries leveraging the Media Source Extensions API set as defined by the W3C. It is an open-source community project, to which we would like to contribute.
Master

Dash.js (github.com/Dash-Industry-Forum/dash.js) is a production quality multimedia player, developed as an initiative of the DASH Industry Forum. It can play back MPEG-DASH content using client-side JavaScript libraries leveraging the Media Source Extensions API set as defined by the W3C. It is an open-source community project, to which we would like to contribute.

For more information about the dash.js player: github.com/Dash-Industry-Forum/dash.js/wiki

Goal

Possible contributions include but are not limited to the following:

  • Designing and implementing AI-based adaptive bitrate (ABR) algorithms for increased efficiency and QoE
  • Experiments to reduce the carbon footprint of streaming with dash.js (“green streaming”)
  • Browser-based testing of the player (e.g., runs tests directly in the browser), and functional testing on different platforms (e.g., Smart TVs)
  • Benchmarking dash.js with other state-of-the-art open-source/commercial players (e.g., Shaka, Theo, Brightcove, Bitmovin etc.) through systematic measurements
  • Updating the reference UI (e.g., allowing for dynamic updates and adding more parameters to the settings): reference.dashif.org/dash.js/nightly/samples/dash-if-reference-player/index.html
  • Implementing a new control bar (e.g., in React/Angular,etc.): github.com/Dash-Industry-Forum/dash.js/discussions/3584
  • Automatic URL generation for DASH-IF live simulator (github.com/Dash-Industry-Forum/dash-live-source-simulator), i.e., UI where we can set the parameters and get the correct URL

Learning outcome

  • Multimedia streaming
  • Video engineering
  • Machine learning / AI
  • Performing research with advanced machine learning algorithms and real world applications

Qualifications

  • Hard working
  • Motivated
  • Interested in learning

Supervisors

  • Pål Halvorsen
  • Cise Midoglu
  • Saeed Sabet

Collaboration partners

Fraunhofer FOKUS, DASH Industry Forum

Contact person